/*
* Write a tuple to the outputstream, in the most efficient format possible.
*/
static void
logicalrep_write_tuple(StringInfo out, Relation rel, HeapTuple tuple)
{
TupleDesc desc;
Datum values[MaxTupleAttributeNumber];
bool isnull[MaxTupleAttributeNumber];
int i;
uint16 nliveatts = 0;
desc = RelationGetDescr(rel);
for (i = 0; i < desc->natts; i++)
{
if (TupleDescAttr(desc, i)->attisdropped)
continue;
nliveatts++;
}
pq_sendint(out, nliveatts, 2);
/* try to allocate enough memory from the get-go */
enlargeStringInfo(out, tuple->t_len +
nliveatts * (1 + 4));
heap_deform_tuple(tuple, desc, values, isnull);
/* Write the values */
for (i = 0; i < desc->natts; i++)
{
HeapTuple typtup;
Form_pg_type typclass;
Form_pg_attribute att = TupleDescAttr(desc, i);
char *outputstr;
/* skip dropped columns */
if (att->attisdropped)
continue;
if (isnull[i])
{
pq_sendbyte(out, 'n'); /* null column */
continue;
}
else if (att->attlen == -1 && VARATT_IS_EXTERNAL_ONDISK(values[i]))
{
pq_sendbyte(out, 'u'); /* unchanged toast column */
continue;
}
typtup = SearchSysCache1(TYPEOID, ObjectIdGetDatum(att->atttypid));
if (!HeapTupleIsValid(typtup))
elog(ERROR, "cache lookup failed for type %u", att->atttypid);
typclass = (Form_pg_type) GETSTRUCT(typtup);
pq_sendbyte(out, 't'); /* 'text' data follows */
outputstr = OidOutputFunctionCall(typclass->typoutput, values[i]);
pq_sendcountedtext(out, outputstr, strlen(outputstr), false);
pfree(outputstr);
ReleaseSysCache(typtup);
}
}
/*
* Read tuples in correct sort order from tuplesort, and load them into
* btree leaves.
*/
static void
_bt_load(BTWriteState *wstate, BTSpool *btspool, BTSpool *btspool2)
{
BTPageState *state = NULL;
bool merge = (btspool2 != NULL);
IndexTuple itup,
itup2 = NULL;
bool should_free,
should_free2,
load1;
TupleDesc tupdes = RelationGetDescr(wstate->index);
int i,
keysz = RelationGetNumberOfAttributes(wstate->index);
ScanKey indexScanKey = NULL;
if (merge)
{
/*
* Another BTSpool for dead tuples exists. Now we have to merge
* btspool and btspool2.
*/
/* the preparation of merge */
itup = tuplesort_getindextuple(btspool->sortstate,
true, &should_free);
itup2 = tuplesort_getindextuple(btspool2->sortstate,
true, &should_free2);
indexScanKey = _bt_mkscankey_nodata(wstate->index);
for (;;)
{
load1 = true; /* load BTSpool next ? */
if (itup2 == NULL)
{
if (itup == NULL)
break;
}
else if (itup != NULL)
{
for (i = 1; i <= keysz; i++)
{
ScanKey entry;
Datum attrDatum1,
attrDatum2;
bool isNull1,
isNull2;
int32 compare;
entry = indexScanKey + i - 1;
attrDatum1 = index_getattr(itup, i, tupdes, &isNull1);
attrDatum2 = index_getattr(itup2, i, tupdes, &isNull2);
if (isNull1)
{
if (isNull2)
compare = 0; /* NULL "=" NULL */
else if (entry->sk_flags & SK_BT_NULLS_FIRST)
compare = -1; /* NULL "<" NOT_NULL */
else
compare = 1; /* NULL ">" NOT_NULL */
}
else if (isNull2)
{
if (entry->sk_flags & SK_BT_NULLS_FIRST)
compare = 1; /* NOT_NULL ">" NULL */
else
compare = -1; /* NOT_NULL "<" NULL */
}
else
{
compare =
DatumGetInt32(FunctionCall2Coll(&entry->sk_func,
entry->sk_collation,
attrDatum1,
attrDatum2));
if (entry->sk_flags & SK_BT_DESC)
compare = -compare;
}
if (compare > 0)
{
load1 = false;
break;
}
else if (compare < 0)
break;
}
}
else
load1 = false;
/* When we see first tuple, create first index page */
if (state == NULL)
state = _bt_pagestate(wstate, 0);
if (load1)
{
_bt_buildadd(wstate, state, itup);
if (should_free)
pfree(itup);
itup = tuplesort_getindextuple(btspool->sortstate,
true, &should_free);
}
else
{
_bt_buildadd(wstate, state, itup2);
if (should_free2)
pfree(itup2);
itup2 = tuplesort_getindextuple(btspool2->sortstate,
true, &should_free2);
}
}
_bt_freeskey(indexScanKey);
}
else
{
/* merge is unnecessary */
while ((itup = tuplesort_getindextuple(btspool->sortstate,
true, &should_free)) != NULL)
{
/* When we see first tuple, create first index page */
if (state == NULL)
state = _bt_pagestate(wstate, 0);
_bt_buildadd(wstate, state, itup);
if (should_free)
pfree(itup);
}
}
/* Close down final pages and write the metapage */
_bt_uppershutdown(wstate, state);
/*
* If the index is WAL-logged, we must fsync it down to disk before it's
* safe to commit the transaction. (For a non-WAL-logged index we don't
* care since the index will be uninteresting after a crash anyway.)
*
* It's obvious that we must do this when not WAL-logging the build. It's
* less obvious that we have to do it even if we did WAL-log the index
* pages. The reason is that since we're building outside shared buffers,
* a CHECKPOINT occurring during the build has no way to flush the
* previously written data to disk (indeed it won't know the index even
* exists). A crash later on would replay WAL from the checkpoint,
* therefore it wouldn't replay our earlier WAL entries. If we do not
* fsync those pages here, they might still not be on disk when the crash
* occurs.
*/
if (RelationNeedsWAL(wstate->index))
{
RelationOpenSmgr(wstate->index);
smgrimmedsync(wstate->index->rd_smgr, MAIN_FORKNUM);
}
}
/*
* InsertRule -
* takes the arguments and inserts them as a row into the system
* relation "pg_rewrite"
*/
static Oid
InsertRule(char *rulname,
int evtype,
Oid eventrel_oid,
bool evinstead,
Node *event_qual,
List *action,
bool replace)
{
char *evqual = nodeToString(event_qual, false);
char *actiontree = nodeToString((Node *) action, false);
Datum values[Natts_pg_rewrite];
bool nulls[Natts_pg_rewrite];
bool replaces[Natts_pg_rewrite];
NameData rname;
Relation pg_rewrite_desc;
HeapTuple tup,
oldtup;
Oid rewriteObjectId;
ObjectAddress myself,
referenced;
bool is_update = false;
/*
* Set up *nulls and *values arrays
*/
MemSet(nulls, false, sizeof(nulls));
namestrcpy(&rname, rulname);
values[Anum_pg_rewrite_rulename - 1] = NameGetDatum(&rname);
values[Anum_pg_rewrite_ev_class - 1] = ObjectIdGetDatum(eventrel_oid);
values[Anum_pg_rewrite_ev_type - 1] = CharGetDatum(evtype + '0');
values[Anum_pg_rewrite_ev_enabled - 1] = CharGetDatum(RULE_FIRES_ON_ORIGIN);
values[Anum_pg_rewrite_is_instead - 1] = BoolGetDatum(evinstead);
values[Anum_pg_rewrite_ev_qual - 1] = CStringGetTextDatum(evqual);
values[Anum_pg_rewrite_ev_action - 1] = CStringGetTextDatum(actiontree);
/*
* Ready to store new pg_rewrite tuple
*/
pg_rewrite_desc = heap_open(RewriteRelationId, RowExclusiveLock);
/*
* Check to see if we are replacing an existing tuple
*/
oldtup = SearchSysCache2(RULERELNAME,
ObjectIdGetDatum(eventrel_oid),
PointerGetDatum(rulname));
if (HeapTupleIsValid(oldtup))
{
if (!replace)
ereport(ERROR,
(errcode(ERRCODE_DUPLICATE_OBJECT),
errmsg("rule \"%s\" for relation \"%s\" already exists",
rulname, get_rel_name(eventrel_oid))));
/*
* When replacing, we don't need to replace every attribute
*/
MemSet(replaces, false, sizeof(replaces));
replaces[Anum_pg_rewrite_ev_type - 1] = true;
replaces[Anum_pg_rewrite_is_instead - 1] = true;
replaces[Anum_pg_rewrite_ev_qual - 1] = true;
replaces[Anum_pg_rewrite_ev_action - 1] = true;
tup = heap_modify_tuple(oldtup, RelationGetDescr(pg_rewrite_desc),
values, nulls, replaces);
simple_heap_update(pg_rewrite_desc, &tup->t_self, tup);
ReleaseSysCache(oldtup);
rewriteObjectId = HeapTupleGetOid(tup);
is_update = true;
}
else
{
tup = heap_form_tuple(pg_rewrite_desc->rd_att, values, nulls);
rewriteObjectId = simple_heap_insert(pg_rewrite_desc, tup);
}
/* Need to update indexes in either case */
CatalogUpdateIndexes(pg_rewrite_desc, tup);
heap_freetuple(tup);
/* If replacing, get rid of old dependencies and make new ones */
if (is_update)
deleteDependencyRecordsFor(RewriteRelationId, rewriteObjectId, false);
/*
* Install dependency on rule's relation to ensure it will go away on
* relation deletion. If the rule is ON SELECT, make the dependency
* implicit --- this prevents deleting a view's SELECT rule. Other kinds
* of rules can be AUTO.
*/
myself.classId = RewriteRelationId;
myself.objectId = rewriteObjectId;
myself.objectSubId = 0;
referenced.classId = RelationRelationId;
referenced.objectId = eventrel_oid;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced,
(evtype == CMD_SELECT) ? DEPENDENCY_INTERNAL : DEPENDENCY_AUTO);
/*
* Also install dependencies on objects referenced in action and qual.
*/
recordDependencyOnExpr(&myself, (Node *) action, NIL,
DEPENDENCY_NORMAL);
if (event_qual != NULL)
{
/* Find query containing OLD/NEW rtable entries */
Query *qry = (Query *) linitial(action);
qry = getInsertSelectQuery(qry, NULL);
recordDependencyOnExpr(&myself, event_qual, qry->rtable,
DEPENDENCY_NORMAL);
}
/* Post creation hook for new rule */
InvokeObjectPostCreateHook(RewriteRelationId, rewriteObjectId, 0);
heap_close(pg_rewrite_desc, RowExclusiveLock);
return rewriteObjectId;
}
/*
* Look to see if we have template information for the given language name.
*/
static PLTemplate *
find_language_template(const char *languageName)
{
PLTemplate *result;
Relation rel;
SysScanDesc scan;
ScanKeyData key;
HeapTuple tup;
rel = heap_open(PLTemplateRelationId, AccessShareLock);
ScanKeyInit(&key,
Anum_pg_pltemplate_tmplname,
BTEqualStrategyNumber, F_NAMEEQ,
CStringGetDatum(languageName));
scan = systable_beginscan(rel, PLTemplateNameIndexId, true,
NULL, 1, &key);
tup = systable_getnext(scan);
if (HeapTupleIsValid(tup))
{
Form_pg_pltemplate tmpl = (Form_pg_pltemplate) GETSTRUCT(tup);
Datum datum;
bool isnull;
result = (PLTemplate *) palloc0(sizeof(PLTemplate));
result->tmpltrusted = tmpl->tmpltrusted;
result->tmpldbacreate = tmpl->tmpldbacreate;
/* Remaining fields are variable-width so we need heap_getattr */
datum = heap_getattr(tup, Anum_pg_pltemplate_tmplhandler,
RelationGetDescr(rel), &isnull);
if (!isnull)
result->tmplhandler = TextDatumGetCString(datum);
datum = heap_getattr(tup, Anum_pg_pltemplate_tmplinline,
RelationGetDescr(rel), &isnull);
if (!isnull)
result->tmplinline = TextDatumGetCString(datum);
datum = heap_getattr(tup, Anum_pg_pltemplate_tmplvalidator,
RelationGetDescr(rel), &isnull);
if (!isnull)
result->tmplvalidator = TextDatumGetCString(datum);
datum = heap_getattr(tup, Anum_pg_pltemplate_tmpllibrary,
RelationGetDescr(rel), &isnull);
if (!isnull)
result->tmpllibrary = TextDatumGetCString(datum);
/* Ignore template if handler or library info is missing */
if (!result->tmplhandler || !result->tmpllibrary)
result = NULL;
}
else
result = NULL;
systable_endscan(scan);
heap_close(rel, AccessShareLock);
return result;
}
/* ----------------------------------------------------------------
* TypeCreate
*
* This does all the necessary work needed to define a new type.
*
* Returns the OID assigned to the new type. If newTypeOid is
* zero (the normal case), a new OID is created; otherwise we
* use exactly that OID.
* ----------------------------------------------------------------
*/
Oid
TypeCreate(Oid newTypeOid,
const char *typeName,
Oid typeNamespace,
Oid relationOid, /* only for relation rowtypes */
char relationKind, /* ditto */
Oid ownerId,
int16 internalSize,
char typeType,
char typeCategory,
bool typePreferred,
char typDelim,
Oid inputProcedure,
Oid outputProcedure,
Oid receiveProcedure,
Oid sendProcedure,
Oid typmodinProcedure,
Oid typmodoutProcedure,
Oid analyzeProcedure,
Oid elementType,
bool isImplicitArray,
Oid arrayType,
Oid baseType,
const char *defaultTypeValue, /* human readable rep */
char *defaultTypeBin, /* cooked rep */
bool passedByValue,
char alignment,
char storage,
int32 typeMod,
int32 typNDims, /* Array dimensions for baseType */
bool typeNotNull)
{
Relation pg_type_desc;
Oid typeObjectId;
bool rebuildDeps = false;
HeapTuple tup;
bool nulls[Natts_pg_type];
bool replaces[Natts_pg_type];
Datum values[Natts_pg_type];
NameData name;
int i;
/*
* We assume that the caller validated the arguments individually, but did
* not check for bad combinations.
*
* Validate size specifications: either positive (fixed-length) or -1
* (varlena) or -2 (cstring).
*/
if (!(internalSize > 0 ||
internalSize == -1 ||
internalSize == -2))
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("invalid type internal size %d",
internalSize)));
if (passedByValue)
{
/*
* Pass-by-value types must have a fixed length that is one of the
* values supported by fetch_att() and store_att_byval(); and the
* alignment had better agree, too. All this code must match
* access/tupmacs.h!
*/
if (internalSize == (int16) sizeof(char))
{
if (alignment != 'c')
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("alignment \"%c\" is invalid for passed-by-value type of size %d",
alignment, internalSize)));
}
else if (internalSize == (int16) sizeof(int16))
{
if (alignment != 's')
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("alignment \"%c\" is invalid for passed-by-value type of size %d",
alignment, internalSize)));
}
else if (internalSize == (int16) sizeof(int32))
{
if (alignment != 'i')
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("alignment \"%c\" is invalid for passed-by-value type of size %d",
alignment, internalSize)));
}
#if SIZEOF_DATUM == 8
else if (internalSize == (int16) sizeof(Datum))
{
if (alignment != 'd')
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("alignment \"%c\" is invalid for passed-by-value type of size %d",
alignment, internalSize)));
}
#endif
else
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("internal size %d is invalid for passed-by-value type",
internalSize)));
}
else
{
/* varlena types must have int align or better */
if (internalSize == -1 && !(alignment == 'i' || alignment == 'd'))
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("alignment \"%c\" is invalid for variable-length type",
alignment)));
/* cstring must have char alignment */
if (internalSize == -2 && !(alignment == 'c'))
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("alignment \"%c\" is invalid for variable-length type",
alignment)));
}
/* Only varlena types can be toasted */
if (storage != 'p' && internalSize != -1)
ereport(ERROR,
(errcode(ERRCODE_INVALID_OBJECT_DEFINITION),
errmsg("fixed-size types must have storage PLAIN")));
/*
* initialize arrays needed for heap_form_tuple or heap_modify_tuple
*/
for (i = 0; i < Natts_pg_type; ++i)
{
nulls[i] = false;
replaces[i] = true;
values[i] = (Datum) 0;
}
/*
* initialize the *values information
*/
i = 0;
namestrcpy(&name, typeName);
values[i++] = NameGetDatum(&name); /* typname */
values[i++] = ObjectIdGetDatum(typeNamespace); /* typnamespace */
values[i++] = ObjectIdGetDatum(ownerId); /* typowner */
values[i++] = Int16GetDatum(internalSize); /* typlen */
values[i++] = BoolGetDatum(passedByValue); /* typbyval */
values[i++] = CharGetDatum(typeType); /* typtype */
values[i++] = CharGetDatum(typeCategory); /* typcategory */
values[i++] = BoolGetDatum(typePreferred); /* typispreferred */
values[i++] = BoolGetDatum(true); /* typisdefined */
values[i++] = CharGetDatum(typDelim); /* typdelim */
values[i++] = ObjectIdGetDatum(relationOid); /* typrelid */
values[i++] = ObjectIdGetDatum(elementType); /* typelem */
values[i++] = ObjectIdGetDatum(arrayType); /* typarray */
values[i++] = ObjectIdGetDatum(inputProcedure); /* typinput */
values[i++] = ObjectIdGetDatum(outputProcedure); /* typoutput */
values[i++] = ObjectIdGetDatum(receiveProcedure); /* typreceive */
values[i++] = ObjectIdGetDatum(sendProcedure); /* typsend */
values[i++] = ObjectIdGetDatum(typmodinProcedure); /* typmodin */
values[i++] = ObjectIdGetDatum(typmodoutProcedure); /* typmodout */
values[i++] = ObjectIdGetDatum(analyzeProcedure); /* typanalyze */
values[i++] = CharGetDatum(alignment); /* typalign */
values[i++] = CharGetDatum(storage); /* typstorage */
values[i++] = BoolGetDatum(typeNotNull); /* typnotnull */
values[i++] = ObjectIdGetDatum(baseType); /* typbasetype */
values[i++] = Int32GetDatum(typeMod); /* typtypmod */
values[i++] = Int32GetDatum(typNDims); /* typndims */
/*
* initialize the default binary value for this type. Check for nulls of
* course.
*/
if (defaultTypeBin)
values[i] = CStringGetTextDatum(defaultTypeBin);
else
nulls[i] = true;
i++; /* typdefaultbin */
/*
* initialize the default value for this type.
*/
if (defaultTypeValue)
values[i] = CStringGetTextDatum(defaultTypeValue);
else
nulls[i] = true;
i++; /* typdefault */
/*
* open pg_type and prepare to insert or update a row.
*
* NOTE: updating will not work correctly in bootstrap mode; but we don't
* expect to be overwriting any shell types in bootstrap mode.
*/
pg_type_desc = heap_open(TypeRelationId, RowExclusiveLock);
tup = SearchSysCacheCopy2(TYPENAMENSP,
CStringGetDatum(typeName),
ObjectIdGetDatum(typeNamespace));
if (HeapTupleIsValid(tup))
{
/*
* check that the type is not already defined. It may exist as a
* shell type, however.
*/
if (((Form_pg_type) GETSTRUCT(tup))->typisdefined)
ereport(ERROR,
(errcode(ERRCODE_DUPLICATE_OBJECT),
errmsg("type \"%s\" already exists", typeName)));
/*
* shell type must have been created by same owner
*/
if (((Form_pg_type) GETSTRUCT(tup))->typowner != ownerId)
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_TYPE, typeName);
/* trouble if caller wanted to force the OID */
if (OidIsValid(newTypeOid))
elog(ERROR, "cannot assign new OID to existing shell type");
/*
* Okay to update existing shell type tuple
*/
tup = heap_modify_tuple(tup,
RelationGetDescr(pg_type_desc),
values,
nulls,
replaces);
simple_heap_update(pg_type_desc, &tup->t_self, tup);
typeObjectId = HeapTupleGetOid(tup);
rebuildDeps = true; /* get rid of shell type's dependencies */
}
else
{
tup = heap_form_tuple(RelationGetDescr(pg_type_desc),
values,
nulls);
/* Force the OID if requested by caller */
if (OidIsValid(newTypeOid))
HeapTupleSetOid(tup, newTypeOid);
else if (OidIsValid(binary_upgrade_next_pg_type_oid))
{
HeapTupleSetOid(tup, binary_upgrade_next_pg_type_oid);
binary_upgrade_next_pg_type_oid = InvalidOid;
}
/* else allow system to assign oid */
typeObjectId = simple_heap_insert(pg_type_desc, tup);
}
/* Update indexes */
CatalogUpdateIndexes(pg_type_desc, tup);
/*
* Create dependencies. We can/must skip this in bootstrap mode.
*/
if (!IsBootstrapProcessingMode())
GenerateTypeDependencies(typeNamespace,
typeObjectId,
relationOid,
relationKind,
ownerId,
inputProcedure,
outputProcedure,
receiveProcedure,
sendProcedure,
typmodinProcedure,
typmodoutProcedure,
analyzeProcedure,
elementType,
isImplicitArray,
baseType,
(defaultTypeBin ?
stringToNode(defaultTypeBin) :
NULL),
rebuildDeps);
/* Post creation hook for new type */
InvokeObjectAccessHook(OAT_POST_CREATE, TypeRelationId, typeObjectId, 0);
/*
* finish up
*/
heap_close(pg_type_desc, RowExclusiveLock);
return typeObjectId;
}
/*
* Alter table space options
*/
void
AlterTableSpaceOptions(AlterTableSpaceOptionsStmt *stmt)
{
Relation rel;
ScanKeyData entry[1];
HeapScanDesc scandesc;
HeapTuple tup;
Datum datum;
Datum newOptions;
Datum repl_val[Natts_pg_tablespace];
bool isnull;
bool repl_null[Natts_pg_tablespace];
bool repl_repl[Natts_pg_tablespace];
HeapTuple newtuple;
/* Search pg_tablespace */
rel = heap_open(TableSpaceRelationId, RowExclusiveLock);
ScanKeyInit(&entry[0],
Anum_pg_tablespace_spcname,
BTEqualStrategyNumber, F_NAMEEQ,
CStringGetDatum(stmt->tablespacename));
scandesc = heap_beginscan(rel, SnapshotNow, 1, entry);
tup = heap_getnext(scandesc, ForwardScanDirection);
if (!HeapTupleIsValid(tup))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("tablespace \"%s\" does not exist",
stmt->tablespacename)));
/* Must be owner of the existing object */
if (!pg_tablespace_ownercheck(HeapTupleGetOid(tup), GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_TABLESPACE,
stmt->tablespacename);
/* Generate new proposed spcoptions (text array) */
datum = heap_getattr(tup, Anum_pg_tablespace_spcoptions,
RelationGetDescr(rel), &isnull);
newOptions = transformRelOptions(isnull ? (Datum) 0 : datum,
stmt->options, NULL, NULL, false,
stmt->isReset);
(void) tablespace_reloptions(newOptions, true);
/* Build new tuple. */
memset(repl_null, false, sizeof(repl_null));
memset(repl_repl, false, sizeof(repl_repl));
if (newOptions != (Datum) 0)
repl_val[Anum_pg_tablespace_spcoptions - 1] = newOptions;
else
repl_null[Anum_pg_tablespace_spcoptions - 1] = true;
repl_repl[Anum_pg_tablespace_spcoptions - 1] = true;
newtuple = heap_modify_tuple(tup, RelationGetDescr(rel), repl_val,
repl_null, repl_repl);
/* Update system catalog. */
simple_heap_update(rel, &newtuple->t_self, newtuple);
CatalogUpdateIndexes(rel, newtuple);
heap_freetuple(newtuple);
/* Conclude heap scan. */
heap_endscan(scandesc);
heap_close(rel, NoLock);
}
/*----------
* _bt_compare() -- Compare scankey to a particular tuple on the page.
*
* The passed scankey must be an insertion-type scankey (see nbtree/README),
* but it can omit the rightmost column(s) of the index.
*
* keysz: number of key conditions to be checked (might be less than the
* number of index columns!)
* page/offnum: location of btree item to be compared to.
*
* This routine returns:
* <0 if scankey < tuple at offnum;
* 0 if scankey == tuple at offnum;
* >0 if scankey > tuple at offnum.
* NULLs in the keys are treated as sortable values. Therefore
* "equality" does not necessarily mean that the item should be
* returned to the caller as a matching key!
*
* CRUCIAL NOTE: on a non-leaf page, the first data key is assumed to be
* "minus infinity": this routine will always claim it is less than the
* scankey. The actual key value stored (if any, which there probably isn't)
* does not matter. This convention allows us to implement the Lehman and
* Yao convention that the first down-link pointer is before the first key.
* See backend/access/nbtree/README for details.
*----------
*/
int32
_bt_compare(Relation rel,
int keysz,
ScanKey scankey,
Page page,
OffsetNumber offnum)
{
TupleDesc itupdesc = RelationGetDescr(rel);
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
IndexTuple itup;
int i;
/*
* Force result ">" if target item is first data item on an internal page
* --- see NOTE above.
*/
if (!P_ISLEAF(opaque) && offnum == P_FIRSTDATAKEY(opaque))
return 1;
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
/*
* The scan key is set up with the attribute number associated with each
* term in the key. It is important that, if the index is multi-key, the
* scan contain the first k key attributes, and that they be in order. If
* you think about how multi-key ordering works, you'll understand why
* this is.
*
* We don't test for violation of this condition here, however. The
* initial setup for the index scan had better have gotten it right (see
* _bt_first).
*/
for (i = 1; i <= keysz; i++)
{
Datum datum;
bool isNull;
int32 result;
datum = index_getattr(itup, scankey->sk_attno, itupdesc, &isNull);
/* see comments about NULLs handling in btbuild */
if (scankey->sk_flags & SK_ISNULL) /* key is NULL */
{
if (isNull)
result = 0; /* NULL "=" NULL */
else if (scankey->sk_flags & SK_BT_NULLS_FIRST)
result = -1; /* NULL "<" NOT_NULL */
else
result = 1; /* NULL ">" NOT_NULL */
}
else if (isNull) /* key is NOT_NULL and item is NULL */
{
if (scankey->sk_flags & SK_BT_NULLS_FIRST)
result = 1; /* NOT_NULL ">" NULL */
else
result = -1; /* NOT_NULL "<" NULL */
}
else
{
/*
* The sk_func needs to be passed the index value as left arg and
* the sk_argument as right arg (they might be of different
* types). Since it is convenient for callers to think of
* _bt_compare as comparing the scankey to the index item, we have
* to flip the sign of the comparison result. (Unless it's a DESC
* column, in which case we *don't* flip the sign.)
*/
result = DatumGetInt32(FunctionCall2(&scankey->sk_func,
datum,
scankey->sk_argument));
if (!(scankey->sk_flags & SK_BT_DESC))
result = -result;
}
/* if the keys are unequal, return the difference */
if (result != 0)
return result;
scankey++;
}
/* if we get here, the keys are equal */
return 0;
}
/*
* pg_get_tableschemadef_string returns the definition of a given table. This
* definition includes table's schema, default column values, not null and check
* constraints. The definition does not include constraints that trigger index
* creations; specifically, unique and primary key constraints are excluded.
*/
static char *
pg_shard_get_tableschemadef_string(Oid tableRelationId)
{
Relation relation = NULL;
char *relationName = NULL;
char relationKind = 0;
TupleDesc tupleDescriptor = NULL;
TupleConstr *tupleConstraints = NULL;
int attributeIndex = 0;
bool firstAttributePrinted = false;
AttrNumber defaultValueIndex = 0;
AttrNumber constraintIndex = 0;
AttrNumber constraintCount = 0;
StringInfoData buffer = { NULL, 0, 0, 0 };
/*
* Instead of retrieving values from system catalogs as other functions in
* ruleutils.c do, we follow an unusual approach here: we open the relation,
* and fetch the relation's tuple descriptor. We do this because the tuple
* descriptor already contains information harnessed from pg_attrdef,
* pg_attribute, pg_constraint, and pg_class; and therefore using the
* descriptor saves us from a lot of additional work.
*/
relation = relation_open(tableRelationId, AccessShareLock);
relationName = generate_relation_name(tableRelationId);
relationKind = relation->rd_rel->relkind;
if (relationKind != RELKIND_RELATION && relationKind != RELKIND_FOREIGN_TABLE)
{
ereport(ERROR, (errcode(ERRCODE_WRONG_OBJECT_TYPE),
errmsg("%s is not a regular or foreign table", relationName)));
}
initStringInfo(&buffer);
if (relationKind == RELKIND_RELATION)
{
appendStringInfo(&buffer, "CREATE TABLE %s (", relationName);
}
else
{
appendStringInfo(&buffer, "CREATE FOREIGN TABLE %s (", relationName);
}
/*
* Iterate over the table's columns. If a particular column is not dropped
* and is not inherited from another table, print the column's name and its
* formatted type.
*/
tupleDescriptor = RelationGetDescr(relation);
tupleConstraints = tupleDescriptor->constr;
for (attributeIndex = 0; attributeIndex < tupleDescriptor->natts; attributeIndex++)
{
Form_pg_attribute attributeForm = tupleDescriptor->attrs[attributeIndex];
if (!attributeForm->attisdropped && attributeForm->attinhcount == 0)
{
const char *attributeName = NULL;
const char *attributeTypeName = NULL;
if (firstAttributePrinted)
{
appendStringInfoString(&buffer, ", ");
}
firstAttributePrinted = true;
attributeName = NameStr(attributeForm->attname);
appendStringInfo(&buffer, "%s ", quote_identifier(attributeName));
attributeTypeName = format_type_with_typemod(attributeForm->atttypid,
attributeForm->atttypmod);
appendStringInfoString(&buffer, attributeTypeName);
/* if this column has a default value, append the default value */
if (attributeForm->atthasdef)
{
AttrDefault *defaultValueList = NULL;
AttrDefault *defaultValue = NULL;
Node *defaultNode = NULL;
List *defaultContext = NULL;
char *defaultString = NULL;
Assert(tupleConstraints != NULL);
defaultValueList = tupleConstraints->defval;
Assert(defaultValueList != NULL);
defaultValue = &(defaultValueList[defaultValueIndex]);
defaultValueIndex++;
Assert(defaultValue->adnum == (attributeIndex + 1));
Assert(defaultValueIndex <= tupleConstraints->num_defval);
/* convert expression to node tree, and prepare deparse context */
defaultNode = (Node *) stringToNode(defaultValue->adbin);
defaultContext = deparse_context_for(relationName, tableRelationId);
/* deparse default value string */
defaultString = deparse_expression(defaultNode, defaultContext,
false, false);
appendStringInfo(&buffer, " DEFAULT %s", defaultString);
}
/* if this column has a not null constraint, append the constraint */
if (attributeForm->attnotnull)
{
appendStringInfoString(&buffer, " NOT NULL");
}
}
}
/*
* Now check if the table has any constraints. If it does, set the number of
* check constraints here. Then iterate over all check constraints and print
* them.
*/
if (tupleConstraints != NULL)
{
constraintCount = tupleConstraints->num_check;
}
for (constraintIndex = 0; constraintIndex < constraintCount; constraintIndex++)
{
ConstrCheck *checkConstraintList = tupleConstraints->check;
ConstrCheck *checkConstraint = &(checkConstraintList[constraintIndex]);
Node *checkNode = NULL;
List *checkContext = NULL;
char *checkString = NULL;
/* if an attribute or constraint has been printed, format properly */
if (firstAttributePrinted || constraintIndex > 0)
{
appendStringInfoString(&buffer, ", ");
}
appendStringInfo(&buffer, "CONSTRAINT %s CHECK ",
quote_identifier(checkConstraint->ccname));
/* convert expression to node tree, and prepare deparse context */
checkNode = (Node *) stringToNode(checkConstraint->ccbin);
checkContext = deparse_context_for(relationName, tableRelationId);
/* deparse check constraint string */
checkString = deparse_expression(checkNode, checkContext, false, false);
appendStringInfoString(&buffer, checkString);
}
/* close create table's outer parentheses */
appendStringInfoString(&buffer, ")");
/*
* If the relation is a foreign table, append the server name and options to
* the create table statement.
*/
if (relationKind == RELKIND_FOREIGN_TABLE)
{
ForeignTable *foreignTable = GetForeignTable(tableRelationId);
ForeignServer *foreignServer = GetForeignServer(foreignTable->serverid);
char *serverName = foreignServer->servername;
appendStringInfo(&buffer, " SERVER %s", quote_identifier(serverName));
AppendOptionListToString(&buffer, foreignTable->options);
}
relation_close(relation, AccessShareLock);
return (buffer.data);
}
/*
* _bitmap_init_buildstate() -- initialize the build state before building
* a bitmap index.
*/
void
_bitmap_init_buildstate(Relation index, BMBuildState *bmstate)
{
MIRROREDLOCK_BUFMGR_DECLARE;
BMMetaPage mp;
HASHCTL hash_ctl;
int hash_flags;
int i;
Buffer metabuf;
/* initialize the build state */
bmstate->bm_tupDesc = RelationGetDescr(index);
bmstate->bm_tidLocsBuffer = (BMTidBuildBuf *)
palloc(sizeof(BMTidBuildBuf));
bmstate->bm_tidLocsBuffer->byte_size = 0;
bmstate->bm_tidLocsBuffer->lov_blocks = NIL;
bmstate->bm_tidLocsBuffer->max_lov_block = InvalidBlockNumber;
// -------- MirroredLock ----------
MIRROREDLOCK_BUFMGR_LOCK;
metabuf = _bitmap_getbuf(index, BM_METAPAGE, BM_READ);
mp = _bitmap_get_metapage_data(index, metabuf);
_bitmap_open_lov_heapandindex(index, mp, &(bmstate->bm_lov_heap),
&(bmstate->bm_lov_index),
RowExclusiveLock);
_bitmap_relbuf(metabuf);
MIRROREDLOCK_BUFMGR_UNLOCK;
// -------- MirroredLock ----------
cur_bmbuild = (BMBuildHashData *)palloc(sizeof(BMBuildHashData));
cur_bmbuild->hash_funcs = (FmgrInfo *)
palloc(sizeof(FmgrInfo) * bmstate->bm_tupDesc->natts);
cur_bmbuild->eq_funcs = (FmgrInfo *)
palloc(sizeof(FmgrInfo) * bmstate->bm_tupDesc->natts);
cur_bmbuild->hash_func_is_strict = (bool *)
palloc(sizeof(bool) * bmstate->bm_tupDesc->natts);
for (i = 0; i < bmstate->bm_tupDesc->natts; i++)
{
Oid typid = bmstate->bm_tupDesc->attrs[i]->atttypid;
Operator optup;
Oid eq_opr;
Oid eq_function;
Oid left_hash_function;
Oid right_hash_function;
optup = equality_oper(typid, false);
eq_opr = oprid(optup);
eq_function = oprfuncid(optup);
ReleaseOperator(optup);
if (!get_op_hash_functions(eq_opr,
&left_hash_function,
&right_hash_function))
{
pfree(cur_bmbuild);
cur_bmbuild = NULL;
break;
}
Assert(left_hash_function == right_hash_function);
fmgr_info(eq_function, &cur_bmbuild->eq_funcs[i]);
fmgr_info(right_hash_function, &cur_bmbuild->hash_funcs[i]);
cur_bmbuild->hash_func_is_strict[i] = func_strict(right_hash_function);
}
if (cur_bmbuild)
{
cur_bmbuild->natts = bmstate->bm_tupDesc->natts;
cur_bmbuild->tmpcxt = AllocSetContextCreate(CurrentMemoryContext,
"Bitmap build temp space",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
/* setup the hash table */
MemSet(&hash_ctl, 0, sizeof(hash_ctl));
/**
* Reserve enough space for the hash key header and then the data segments (values followed by nulls)
*/
hash_ctl.keysize = MAXALIGN(sizeof(BMBuildHashKey)) +
MAXALIGN(sizeof(Datum) * cur_bmbuild->natts) +
MAXALIGN(sizeof(bool) * cur_bmbuild->natts);
hash_ctl.entrysize = hash_ctl.keysize + sizeof(BMBuildLovData) + 200;
hash_ctl.hash = build_hash_key;
hash_ctl.match = build_match_key;
hash_ctl.keycopy = build_keycopy;
hash_ctl.hcxt = AllocSetContextCreate(CurrentMemoryContext,
"Bitmap build hash table",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
cur_bmbuild->hash_cxt = hash_ctl.hcxt;
hash_flags = HASH_ELEM | HASH_FUNCTION | HASH_COMPARE | HASH_CONTEXT | HASH_KEYCOPY;
bmstate->lovitem_hash = hash_create("Bitmap index build lov item hash",
100, &hash_ctl, hash_flags);
bmstate->lovitem_hashKeySize = hash_ctl.keysize;
}
else
{
int attno;
bmstate->lovitem_hash = NULL;
bmstate->lovitem_hashKeySize = 0;
bmstate->bm_lov_scanKeys =
(ScanKey)palloc0(bmstate->bm_tupDesc->natts * sizeof(ScanKeyData));
for (attno = 0; attno < bmstate->bm_tupDesc->natts; attno++)
{
RegProcedure opfuncid;
Oid atttypid;
atttypid = bmstate->bm_tupDesc->attrs[attno]->atttypid;
opfuncid = equality_oper_funcid(atttypid);
ScanKeyEntryInitialize(&(bmstate->bm_lov_scanKeys[attno]), SK_ISNULL,
attno + 1, BTEqualStrategyNumber, InvalidOid,
opfuncid, 0);
}
bmstate->bm_lov_scanDesc = index_beginscan(bmstate->bm_lov_heap,
bmstate->bm_lov_index, ActiveSnapshot,
bmstate->bm_tupDesc->natts,
bmstate->bm_lov_scanKeys);
}
/*
* We need to log index creation in WAL iff WAL archiving is enabled
* AND it's not a temp index. Currently, since building an index
* writes page to the shared buffer, we can't disable WAL archiving.
* We will add this shortly.
*/
bmstate->use_wal = !XLog_UnconvertedCanBypassWal() && !index->rd_istemp;
}
/*
* shdepChangeDep
*
* Update shared dependency records to account for an updated referenced
* object. This is an internal workhorse for operations such as changing
* an object's owner.
*
* There must be no more than one existing entry for the given dependent
* object and dependency type! So in practice this can only be used for
* updating SHARED_DEPENDENCY_OWNER entries, which should have that property.
*
* If there is no previous entry, we assume it was referencing a PINned
* object, so we create a new entry. If the new referenced object is
* PINned, we don't create an entry (and drop the old one, if any).
*
* sdepRel must be the pg_shdepend relation, already opened and suitably
* locked.
*/
static void
shdepChangeDep(Relation sdepRel,
Oid classid, Oid objid, int32 objsubid,
Oid refclassid, Oid refobjid,
SharedDependencyType deptype)
{
Oid dbid = classIdGetDbId(classid);
HeapTuple oldtup = NULL;
HeapTuple scantup;
ScanKeyData key[4];
SysScanDesc scan;
/*
* Make sure the new referenced object doesn't go away while we record the
* dependency.
*/
shdepLockAndCheckObject(refclassid, refobjid);
/*
* Look for a previous entry
*/
ScanKeyInit(&key[0],
Anum_pg_shdepend_dbid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(dbid));
ScanKeyInit(&key[1],
Anum_pg_shdepend_classid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(classid));
ScanKeyInit(&key[2],
Anum_pg_shdepend_objid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(objid));
ScanKeyInit(&key[3],
Anum_pg_shdepend_objsubid,
BTEqualStrategyNumber, F_INT4EQ,
Int32GetDatum(objsubid));
scan = systable_beginscan(sdepRel, SharedDependDependerIndexId, true,
NULL, 4, key);
while ((scantup = systable_getnext(scan)) != NULL)
{
/* Ignore if not of the target dependency type */
if (((Form_pg_shdepend) GETSTRUCT(scantup))->deptype != deptype)
continue;
/* Caller screwed up if multiple matches */
if (oldtup)
elog(ERROR,
"multiple pg_shdepend entries for object %u/%u/%d deptype %c",
classid, objid, objsubid, deptype);
oldtup = heap_copytuple(scantup);
}
systable_endscan(scan);
if (isSharedObjectPinned(refclassid, refobjid, sdepRel))
{
/* No new entry needed, so just delete existing entry if any */
if (oldtup)
simple_heap_delete(sdepRel, &oldtup->t_self);
}
else if (oldtup)
{
/* Need to update existing entry */
Form_pg_shdepend shForm = (Form_pg_shdepend) GETSTRUCT(oldtup);
/* Since oldtup is a copy, we can just modify it in-memory */
shForm->refclassid = refclassid;
shForm->refobjid = refobjid;
simple_heap_update(sdepRel, &oldtup->t_self, oldtup);
/* keep indexes current */
CatalogUpdateIndexes(sdepRel, oldtup);
}
else
{
/* Need to insert new entry */
Datum values[Natts_pg_shdepend];
bool nulls[Natts_pg_shdepend];
memset(nulls, false, sizeof(nulls));
values[Anum_pg_shdepend_dbid - 1] = ObjectIdGetDatum(dbid);
values[Anum_pg_shdepend_classid - 1] = ObjectIdGetDatum(classid);
values[Anum_pg_shdepend_objid - 1] = ObjectIdGetDatum(objid);
values[Anum_pg_shdepend_objsubid - 1] = Int32GetDatum(objsubid);
values[Anum_pg_shdepend_refclassid - 1] = ObjectIdGetDatum(refclassid);
values[Anum_pg_shdepend_refobjid - 1] = ObjectIdGetDatum(refobjid);
values[Anum_pg_shdepend_deptype - 1] = CharGetDatum(deptype);
/*
* we are reusing oldtup just to avoid declaring a new variable, but
* it's certainly a new tuple
*/
oldtup = heap_form_tuple(RelationGetDescr(sdepRel), values, nulls);
simple_heap_insert(sdepRel, oldtup);
/* keep indexes current */
CatalogUpdateIndexes(sdepRel, oldtup);
}
if (oldtup)
heap_freetuple(oldtup);
}
/*
* unique_key_recheck - trigger function to do a deferred uniqueness check.
*
* This now also does deferred exclusion-constraint checks, so the name is
* somewhat historical.
*
* This is invoked as an AFTER ROW trigger for both INSERT and UPDATE,
* for any rows recorded as potentially violating a deferrable unique
* or exclusion constraint.
*
* This may be an end-of-statement check, a commit-time check, or a
* check triggered by a SET CONSTRAINTS command.
*/
Datum
unique_key_recheck(PG_FUNCTION_ARGS)
{
TriggerData *trigdata = (TriggerData *) fcinfo->context;
const char *funcname = "unique_key_recheck";
HeapTuple new_row;
ItemPointerData tmptid;
Relation indexRel;
IndexInfo *indexInfo;
EState *estate;
ExprContext *econtext;
TupleTableSlot *slot;
Datum values[INDEX_MAX_KEYS];
bool isnull[INDEX_MAX_KEYS];
/*
* Make sure this is being called as an AFTER ROW trigger. Note:
* translatable error strings are shared with ri_triggers.c, so resist the
* temptation to fold the function name into them.
*/
if (!CALLED_AS_TRIGGER(fcinfo))
ereport(ERROR,
(errcode(ERRCODE_E_R_I_E_TRIGGER_PROTOCOL_VIOLATED),
errmsg("function \"%s\" was not called by trigger manager",
funcname)));
if (!TRIGGER_FIRED_AFTER(trigdata->tg_event) ||
!TRIGGER_FIRED_FOR_ROW(trigdata->tg_event))
ereport(ERROR,
(errcode(ERRCODE_E_R_I_E_TRIGGER_PROTOCOL_VIOLATED),
errmsg("function \"%s\" must be fired AFTER ROW",
funcname)));
/*
* Get the new data that was inserted/updated.
*/
if (TRIGGER_FIRED_BY_INSERT(trigdata->tg_event))
new_row = trigdata->tg_trigtuple;
else if (TRIGGER_FIRED_BY_UPDATE(trigdata->tg_event))
new_row = trigdata->tg_newtuple;
else
{
ereport(ERROR,
(errcode(ERRCODE_E_R_I_E_TRIGGER_PROTOCOL_VIOLATED),
errmsg("function \"%s\" must be fired for INSERT or UPDATE",
funcname)));
new_row = NULL; /* keep compiler quiet */
}
/*
* If the new_row is now dead (ie, inserted and then deleted within our
* transaction), we can skip the check. However, we have to be careful,
* because this trigger gets queued only in response to index insertions;
* which means it does not get queued for HOT updates. The row we are
* called for might now be dead, but have a live HOT child, in which case
* we still need to make the check --- effectively, we're applying the
* check against the live child row, although we can use the values from
* this row since by definition all columns of interest to us are the
* same.
*
* This might look like just an optimization, because the index AM will
* make this identical test before throwing an error. But it's actually
* needed for correctness, because the index AM will also throw an error
* if it doesn't find the index entry for the row. If the row's dead then
* it's possible the index entry has also been marked dead, and even
* removed.
*/
tmptid = new_row->t_self;
if (!heap_hot_search(&tmptid, trigdata->tg_relation, SnapshotSelf, NULL))
{
/*
* All rows in the HOT chain are dead, so skip the check.
*/
return PointerGetDatum(NULL);
}
/*
* Open the index, acquiring a RowExclusiveLock, just as if we were going
* to update it. (This protects against possible changes of the index
* schema, not against concurrent updates.)
*/
indexRel = index_open(trigdata->tg_trigger->tgconstrindid,
RowExclusiveLock);
indexInfo = BuildIndexInfo(indexRel);
/*
* The heap tuple must be put into a slot for FormIndexDatum.
*/
slot = MakeSingleTupleTableSlot(RelationGetDescr(trigdata->tg_relation));
ExecStoreHeapTuple(new_row, slot, InvalidBuffer, false);
/*
* Typically the index won't have expressions, but if it does we need an
* EState to evaluate them. We need it for exclusion constraints too,
* even if they are just on simple columns.
*/
if (indexInfo->ii_Expressions != NIL ||
indexInfo->ii_ExclusionOps != NULL)
{
estate = CreateExecutorState();
econtext = GetPerTupleExprContext(estate);
econtext->ecxt_scantuple = slot;
}
else
estate = NULL;
/*
* Form the index values and isnull flags for the index entry that we need
* to check.
*
* Note: if the index uses functions that are not as immutable as they are
* supposed to be, this could produce an index tuple different from the
* original. The index AM can catch such errors by verifying that it
* finds a matching index entry with the tuple's TID. For exclusion
* constraints we check this in check_exclusion_constraint().
*/
FormIndexDatum(indexInfo, slot, estate, values, isnull);
/*
* Now do the appropriate check.
*/
if (indexInfo->ii_ExclusionOps == NULL)
{
/*
* Note: this is not a real insert; it is a check that the index entry
* that has already been inserted is unique. Passing t_self is
* correct even if t_self is now dead, because that is the TID the
* index will know about.
*/
index_insert(indexRel, values, isnull, &(new_row->t_self),
trigdata->tg_relation, UNIQUE_CHECK_EXISTING);
}
else
{
/*
* For exclusion constraints we just do the normal check, but now it's
* okay to throw error. In the HOT-update case, we must use the live
* HOT child's TID here, else check_exclusion_constraint will think
* the child is a conflict.
*/
check_exclusion_constraint(trigdata->tg_relation, indexRel, indexInfo,
&tmptid, values, isnull,
estate, false, false);
}
/*
* If that worked, then this index entry is unique or non-excluded, and we
* are done.
*/
if (estate != NULL)
FreeExecutorState(estate);
ExecDropSingleTupleTableSlot(slot);
index_close(indexRel, RowExclusiveLock);
return PointerGetDatum(NULL);
}
/*
* Rebuild the event trigger cache.
*/
static void
BuildEventTriggerCache(void)
{
HASHCTL ctl;
HTAB *cache;
MemoryContext oldcontext;
Relation rel;
Relation irel;
SysScanDesc scan;
if (EventTriggerCacheContext != NULL)
{
/*
* Free up any memory already allocated in EventTriggerCacheContext.
* This can happen either because a previous rebuild failed, or
* because an invalidation happened before the rebuild was complete.
*/
MemoryContextResetAndDeleteChildren(EventTriggerCacheContext);
}
else
{
/*
* This is our first time attempting to build the cache, so we need
* to set up the memory context and register a syscache callback to
* capture future invalidation events.
*/
if (CacheMemoryContext == NULL)
CreateCacheMemoryContext();
EventTriggerCacheContext =
AllocSetContextCreate(CacheMemoryContext,
"EventTriggerCache",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
CacheRegisterSyscacheCallback(EVENTTRIGGEROID,
InvalidateEventCacheCallback,
(Datum) 0);
}
/* Switch to correct memory context. */
oldcontext = MemoryContextSwitchTo(EventTriggerCacheContext);
/* Prevent the memory context from being nuked while we're rebuilding. */
EventTriggerCacheState = ETCS_REBUILD_STARTED;
/* Create new hash table. */
MemSet(&ctl, 0, sizeof(ctl));
ctl.keysize = sizeof(EventTriggerEvent);
ctl.entrysize = sizeof(EventTriggerCacheEntry);
ctl.hash = tag_hash;
ctl.hcxt = EventTriggerCacheContext;
cache = hash_create("Event Trigger Cache", 32, &ctl,
HASH_ELEM | HASH_FUNCTION | HASH_CONTEXT);
/*
* Prepare to scan pg_event_trigger in name order. We use an MVCC
* snapshot to avoid getting inconsistent results if the table is
* being concurrently updated.
*/
rel = relation_open(EventTriggerRelationId, AccessShareLock);
irel = index_open(EventTriggerNameIndexId, AccessShareLock);
scan = systable_beginscan_ordered(rel, irel, GetLatestSnapshot(), 0, NULL);
/*
* Build a cache item for each pg_event_trigger tuple, and append each
* one to the appropriate cache entry.
*/
for (;;)
{
HeapTuple tup;
Form_pg_event_trigger form;
char *evtevent;
EventTriggerEvent event;
EventTriggerCacheItem *item;
Datum evttags;
bool evttags_isnull;
EventTriggerCacheEntry *entry;
bool found;
/* Get next tuple. */
tup = systable_getnext_ordered(scan, ForwardScanDirection);
if (!HeapTupleIsValid(tup))
break;
/* Skip trigger if disabled. */
form = (Form_pg_event_trigger) GETSTRUCT(tup);
if (form->evtenabled == TRIGGER_DISABLED)
continue;
/* Decode event name. */
evtevent = NameStr(form->evtevent);
if (strcmp(evtevent, "ddl_command_start") == 0)
event = EVT_DDLCommandStart;
else
continue;
/* Allocate new cache item. */
item = palloc0(sizeof(EventTriggerCacheItem));
item->fnoid = form->evtfoid;
item->enabled = form->evtenabled;
/* Decode and sort tags array. */
evttags = heap_getattr(tup, Anum_pg_event_trigger_evttags,
RelationGetDescr(rel), &evttags_isnull);
if (!evttags_isnull)
{
item->ntags = DecodeTextArrayToCString(evttags, &item->tag);
qsort(item->tag, item->ntags, sizeof(char *), pg_qsort_strcmp);
}
/* Add to cache entry. */
entry = hash_search(cache, &event, HASH_ENTER, &found);
if (found)
entry->triggerlist = lappend(entry->triggerlist, item);
else
entry->triggerlist = list_make1(item);
}
/* Done with pg_event_trigger scan. */
systable_endscan_ordered(scan);
index_close(irel, AccessShareLock);
relation_close(rel, AccessShareLock);
/* Restore previous memory context. */
MemoryContextSwitchTo(oldcontext);
/* Install new cache. */
EventTriggerCache = cache;
/*
* If the cache has been invalidated since we entered this routine, we
* still use and return the cache we just finished constructing, to avoid
* infinite loops, but we leave the cache marked stale so that we'll
* rebuild it again on next access. Otherwise, we mark the cache valid.
*/
if (EventTriggerCacheState == ETCS_REBUILD_STARTED)
EventTriggerCacheState = ETCS_VALID;
}
/* ----------------------------------------------------------------
* ProcedureCreate
*
* Note: allParameterTypes, parameterModes, parameterNames, and proconfig
* are either arrays of the proper types or NULL. We declare them Datum,
* not "ArrayType *", to avoid importing array.h into pg_proc_fn.h.
* ----------------------------------------------------------------
*/
Oid
ProcedureCreate(const char *procedureName,
Oid procNamespace,
bool replace,
bool returnsSet,
Oid returnType,
Oid proowner,
Oid languageObjectId,
Oid languageValidator,
const char *prosrc,
const char *probin,
bool isAgg,
bool isWindowFunc,
bool security_definer,
bool isLeakProof,
bool isStrict,
char volatility,
oidvector *parameterTypes,
Datum allParameterTypes,
Datum parameterModes,
Datum parameterNames,
List *parameterDefaults,
Datum proconfig,
float4 procost,
float4 prorows)
{
Oid retval;
int parameterCount;
int allParamCount;
Oid *allParams;
char *paramModes = NULL;
bool genericInParam = false;
bool genericOutParam = false;
bool anyrangeInParam = false;
bool anyrangeOutParam = false;
bool internalInParam = false;
bool internalOutParam = false;
Oid variadicType = InvalidOid;
Acl *proacl = NULL;
Relation rel;
HeapTuple tup;
HeapTuple oldtup;
bool nulls[Natts_pg_proc];
Datum values[Natts_pg_proc];
bool replaces[Natts_pg_proc];
Oid relid;
NameData procname;
TupleDesc tupDesc;
bool is_update;
ObjectAddress myself,
referenced;
int i;
/*
* sanity checks
*/
Assert(PointerIsValid(prosrc));
parameterCount = parameterTypes->dim1;
if (parameterCount < 0 || parameterCount > FUNC_MAX_ARGS)
ereport(ERROR,
(errcode(ERRCODE_TOO_MANY_ARGUMENTS),
errmsg_plural("functions cannot have more than %d argument",
"functions cannot have more than %d arguments",
FUNC_MAX_ARGS,
FUNC_MAX_ARGS)));
/* note: the above is correct, we do NOT count output arguments */
/* Deconstruct array inputs */
if (allParameterTypes != PointerGetDatum(NULL))
{
/*
* We expect the array to be a 1-D OID array; verify that. We don't
* need to use deconstruct_array() since the array data is just going
* to look like a C array of OID values.
*/
ArrayType *allParamArray = (ArrayType *) DatumGetPointer(allParameterTypes);
allParamCount = ARR_DIMS(allParamArray)[0];
if (ARR_NDIM(allParamArray) != 1 ||
allParamCount <= 0 ||
ARR_HASNULL(allParamArray) ||
ARR_ELEMTYPE(allParamArray) != OIDOID)
elog(ERROR, "allParameterTypes is not a 1-D Oid array");
allParams = (Oid *) ARR_DATA_PTR(allParamArray);
Assert(allParamCount >= parameterCount);
/* we assume caller got the contents right */
}
else
{
allParamCount = parameterCount;
allParams = parameterTypes->values;
}
if (parameterModes != PointerGetDatum(NULL))
{
/*
* We expect the array to be a 1-D CHAR array; verify that. We don't
* need to use deconstruct_array() since the array data is just going
* to look like a C array of char values.
*/
ArrayType *modesArray = (ArrayType *) DatumGetPointer(parameterModes);
if (ARR_NDIM(modesArray) != 1 ||
ARR_DIMS(modesArray)[0] != allParamCount ||
ARR_HASNULL(modesArray) ||
ARR_ELEMTYPE(modesArray) != CHAROID)
elog(ERROR, "parameterModes is not a 1-D char array");
paramModes = (char *) ARR_DATA_PTR(modesArray);
}
/*
* Detect whether we have polymorphic or INTERNAL arguments. The first
* loop checks input arguments, the second output arguments.
*/
for (i = 0; i < parameterCount; i++)
{
switch (parameterTypes->values[i])
{
case ANYARRAYOID:
case ANYELEMENTOID:
case ANYNONARRAYOID:
case ANYENUMOID:
genericInParam = true;
break;
case ANYRANGEOID:
genericInParam = true;
anyrangeInParam = true;
break;
case INTERNALOID:
internalInParam = true;
break;
}
}
if (allParameterTypes != PointerGetDatum(NULL))
{
for (i = 0; i < allParamCount; i++)
{
if (paramModes == NULL ||
paramModes[i] == PROARGMODE_IN ||
paramModes[i] == PROARGMODE_VARIADIC)
continue; /* ignore input-only params */
switch (allParams[i])
{
case ANYARRAYOID:
case ANYELEMENTOID:
case ANYNONARRAYOID:
case ANYENUMOID:
genericOutParam = true;
break;
case ANYRANGEOID:
genericOutParam = true;
anyrangeOutParam = true;
break;
case INTERNALOID:
internalOutParam = true;
break;
}
}
}
/*
* Do not allow polymorphic return type unless at least one input argument
* is polymorphic. ANYRANGE return type is even stricter: must have an
* ANYRANGE input (since we can't deduce the specific range type from
* ANYELEMENT). Also, do not allow return type INTERNAL unless at least
* one input argument is INTERNAL.
*/
if ((IsPolymorphicType(returnType) || genericOutParam)
&& !genericInParam)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot determine result data type"),
errdetail("A function returning a polymorphic type must have at least one polymorphic argument.")));
if ((returnType == ANYRANGEOID || anyrangeOutParam) &&
!anyrangeInParam)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot determine result data type"),
errdetail("A function returning ANYRANGE must have at least one ANYRANGE argument.")));
if ((returnType == INTERNALOID || internalOutParam) && !internalInParam)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("unsafe use of pseudo-type \"internal\""),
errdetail("A function returning \"internal\" must have at least one \"internal\" argument.")));
/*
* don't allow functions of complex types that have the same name as
* existing attributes of the type
*/
if (parameterCount == 1 &&
OidIsValid(parameterTypes->values[0]) &&
(relid = typeidTypeRelid(parameterTypes->values[0])) != InvalidOid &&
get_attnum(relid, procedureName) != InvalidAttrNumber)
ereport(ERROR,
(errcode(ERRCODE_DUPLICATE_COLUMN),
errmsg("\"%s\" is already an attribute of type %s",
procedureName,
format_type_be(parameterTypes->values[0]))));
if (paramModes != NULL)
{
/*
* Only the last input parameter can be variadic; if it is, save its
* element type. Errors here are just elog since caller should have
* checked this already.
*/
for (i = 0; i < allParamCount; i++)
{
switch (paramModes[i])
{
case PROARGMODE_IN:
case PROARGMODE_INOUT:
if (OidIsValid(variadicType))
elog(ERROR, "variadic parameter must be last");
break;
case PROARGMODE_OUT:
case PROARGMODE_TABLE:
/* okay */
break;
case PROARGMODE_VARIADIC:
if (OidIsValid(variadicType))
elog(ERROR, "variadic parameter must be last");
switch (allParams[i])
{
case ANYOID:
variadicType = ANYOID;
break;
case ANYARRAYOID:
variadicType = ANYELEMENTOID;
break;
default:
variadicType = get_element_type(allParams[i]);
if (!OidIsValid(variadicType))
elog(ERROR, "variadic parameter is not an array");
break;
}
break;
default:
elog(ERROR, "invalid parameter mode '%c'", paramModes[i]);
break;
}
}
}
/*
* All seems OK; prepare the data to be inserted into pg_proc.
*/
for (i = 0; i < Natts_pg_proc; ++i)
{
nulls[i] = false;
values[i] = (Datum) 0;
replaces[i] = true;
}
namestrcpy(&procname, procedureName);
values[Anum_pg_proc_proname - 1] = NameGetDatum(&procname);
values[Anum_pg_proc_pronamespace - 1] = ObjectIdGetDatum(procNamespace);
values[Anum_pg_proc_proowner - 1] = ObjectIdGetDatum(proowner);
values[Anum_pg_proc_prolang - 1] = ObjectIdGetDatum(languageObjectId);
values[Anum_pg_proc_procost - 1] = Float4GetDatum(procost);
values[Anum_pg_proc_prorows - 1] = Float4GetDatum(prorows);
values[Anum_pg_proc_provariadic - 1] = ObjectIdGetDatum(variadicType);
values[Anum_pg_proc_protransform - 1] = ObjectIdGetDatum(InvalidOid);
values[Anum_pg_proc_proisagg - 1] = BoolGetDatum(isAgg);
values[Anum_pg_proc_proiswindow - 1] = BoolGetDatum(isWindowFunc);
values[Anum_pg_proc_prosecdef - 1] = BoolGetDatum(security_definer);
values[Anum_pg_proc_proleakproof - 1] = BoolGetDatum(isLeakProof);
values[Anum_pg_proc_proisstrict - 1] = BoolGetDatum(isStrict);
values[Anum_pg_proc_proretset - 1] = BoolGetDatum(returnsSet);
values[Anum_pg_proc_provolatile - 1] = CharGetDatum(volatility);
values[Anum_pg_proc_pronargs - 1] = UInt16GetDatum(parameterCount);
values[Anum_pg_proc_pronargdefaults - 1] = UInt16GetDatum(list_length(parameterDefaults));
values[Anum_pg_proc_prorettype - 1] = ObjectIdGetDatum(returnType);
values[Anum_pg_proc_proargtypes - 1] = PointerGetDatum(parameterTypes);
if (allParameterTypes != PointerGetDatum(NULL))
values[Anum_pg_proc_proallargtypes - 1] = allParameterTypes;
else
nulls[Anum_pg_proc_proallargtypes - 1] = true;
if (parameterModes != PointerGetDatum(NULL))
values[Anum_pg_proc_proargmodes - 1] = parameterModes;
else
nulls[Anum_pg_proc_proargmodes - 1] = true;
if (parameterNames != PointerGetDatum(NULL))
values[Anum_pg_proc_proargnames - 1] = parameterNames;
else
nulls[Anum_pg_proc_proargnames - 1] = true;
if (parameterDefaults != NIL)
values[Anum_pg_proc_proargdefaults - 1] = CStringGetTextDatum(nodeToString(parameterDefaults));
else
nulls[Anum_pg_proc_proargdefaults - 1] = true;
values[Anum_pg_proc_prosrc - 1] = CStringGetTextDatum(prosrc);
if (probin)
values[Anum_pg_proc_probin - 1] = CStringGetTextDatum(probin);
else
nulls[Anum_pg_proc_probin - 1] = true;
if (proconfig != PointerGetDatum(NULL))
values[Anum_pg_proc_proconfig - 1] = proconfig;
else
nulls[Anum_pg_proc_proconfig - 1] = true;
/* proacl will be determined later */
rel = heap_open(ProcedureRelationId, RowExclusiveLock);
tupDesc = RelationGetDescr(rel);
/* Check for pre-existing definition */
oldtup = SearchSysCache3(PROCNAMEARGSNSP,
PointerGetDatum(procedureName),
PointerGetDatum(parameterTypes),
ObjectIdGetDatum(procNamespace));
if (HeapTupleIsValid(oldtup))
{
/* There is one; okay to replace it? */
Form_pg_proc oldproc = (Form_pg_proc) GETSTRUCT(oldtup);
Datum proargnames;
bool isnull;
if (!replace)
ereport(ERROR,
(errcode(ERRCODE_DUPLICATE_FUNCTION),
errmsg("function \"%s\" already exists with same argument types",
procedureName)));
if (!pg_proc_ownercheck(HeapTupleGetOid(oldtup), proowner))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_PROC,
procedureName);
/*
* Not okay to change the return type of the existing proc, since
* existing rules, views, etc may depend on the return type.
*/
if (returnType != oldproc->prorettype ||
returnsSet != oldproc->proretset)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot change return type of existing function"),
errhint("Use DROP FUNCTION first.")));
/*
* If it returns RECORD, check for possible change of record type
* implied by OUT parameters
*/
if (returnType == RECORDOID)
{
TupleDesc olddesc;
TupleDesc newdesc;
olddesc = build_function_result_tupdesc_t(oldtup);
newdesc = build_function_result_tupdesc_d(allParameterTypes,
parameterModes,
parameterNames);
if (olddesc == NULL && newdesc == NULL)
/* ok, both are runtime-defined RECORDs */ ;
else if (olddesc == NULL || newdesc == NULL ||
!equalTupleDescs(olddesc, newdesc))
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot change return type of existing function"),
errdetail("Row type defined by OUT parameters is different."),
errhint("Use DROP FUNCTION first.")));
}
/*
* If there were any named input parameters, check to make sure the
* names have not been changed, as this could break existing calls. We
* allow adding names to formerly unnamed parameters, though.
*/
proargnames = SysCacheGetAttr(PROCNAMEARGSNSP, oldtup,
Anum_pg_proc_proargnames,
&isnull);
if (!isnull)
{
Datum proargmodes;
char **old_arg_names;
char **new_arg_names;
int n_old_arg_names;
int n_new_arg_names;
int j;
proargmodes = SysCacheGetAttr(PROCNAMEARGSNSP, oldtup,
Anum_pg_proc_proargmodes,
&isnull);
if (isnull)
proargmodes = PointerGetDatum(NULL); /* just to be sure */
n_old_arg_names = get_func_input_arg_names(proargnames,
proargmodes,
&old_arg_names);
n_new_arg_names = get_func_input_arg_names(parameterNames,
parameterModes,
&new_arg_names);
for (j = 0; j < n_old_arg_names; j++)
{
if (old_arg_names[j] == NULL)
continue;
if (j >= n_new_arg_names || new_arg_names[j] == NULL ||
strcmp(old_arg_names[j], new_arg_names[j]) != 0)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot change name of input parameter \"%s\"",
old_arg_names[j]),
errhint("Use DROP FUNCTION first.")));
}
}
/*
* If there are existing defaults, check compatibility: redefinition
* must not remove any defaults nor change their types. (Removing a
* default might cause a function to fail to satisfy an existing call.
* Changing type would only be possible if the associated parameter is
* polymorphic, and in such cases a change of default type might alter
* the resolved output type of existing calls.)
*/
if (oldproc->pronargdefaults != 0)
{
Datum proargdefaults;
List *oldDefaults;
ListCell *oldlc;
ListCell *newlc;
if (list_length(parameterDefaults) < oldproc->pronargdefaults)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot remove parameter defaults from existing function"),
errhint("Use DROP FUNCTION first.")));
proargdefaults = SysCacheGetAttr(PROCNAMEARGSNSP, oldtup,
Anum_pg_proc_proargdefaults,
&isnull);
Assert(!isnull);
oldDefaults = (List *) stringToNode(TextDatumGetCString(proargdefaults));
Assert(IsA(oldDefaults, List));
Assert(list_length(oldDefaults) == oldproc->pronargdefaults);
/* new list can have more defaults than old, advance over 'em */
newlc = list_head(parameterDefaults);
for (i = list_length(parameterDefaults) - oldproc->pronargdefaults;
i > 0;
i--)
newlc = lnext(newlc);
foreach(oldlc, oldDefaults)
{
Node *oldDef = (Node *) lfirst(oldlc);
Node *newDef = (Node *) lfirst(newlc);
if (exprType(oldDef) != exprType(newDef))
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("cannot change data type of existing parameter default value"),
errhint("Use DROP FUNCTION first.")));
newlc = lnext(newlc);
}
}
/**
*begin scanning of a parquet relation
*/
ParquetScanDesc
parquet_beginscan(
Relation relation,
Snapshot parquetMetaDataSnapshot,
TupleDesc relationTupleDesc,
bool *proj)
{
ParquetScanDesc scan;
AppendOnlyEntry *aoEntry;
AppendOnlyStorageAttributes *attr;
/*
* increment relation ref count while scanning relation
*
* This is just to make really sure the relcache entry won't go away while
* the scan has a pointer to it. Caller should be holding the rel open
* anyway, so this is redundant in all normal scenarios...
*/
RelationIncrementReferenceCount(relation);
/* allocate scan descriptor */
scan = (ParquetScanDescData *)palloc0(sizeof(ParquetScanDescData));
/*
* Get the pg_appendonly information for this table
*/
aoEntry = GetAppendOnlyEntry(RelationGetRelid(relation), parquetMetaDataSnapshot);
scan->aoEntry = aoEntry;
Assert(aoEntry->majorversion == 1 && aoEntry->minorversion == 0);
#ifdef FAULT_INJECTOR
FaultInjector_InjectFaultIfSet(
FailQeWhenBeginParquetScan,
DDLNotSpecified,
"", // databaseName
""); // tableName
#endif
/*
* initialize the scan descriptor
*/
scan->pqs_filenamepath_maxlen = AOSegmentFilePathNameLen(relation) + 1;
scan->pqs_filenamepath = (char*)palloc0(scan->pqs_filenamepath_maxlen);
scan->pqs_rd = relation;
scan->parquetScanInitContext = CurrentMemoryContext;
/*
* Fill in Parquet Storage layer attributes.
*/
attr = &scan->storageAttributes;
/*
* These attributes describe the AppendOnly format to be scanned.
*/
if (aoEntry->compresstype == NULL || pg_strcasecmp(aoEntry->compresstype, "none") == 0)
attr->compress = false;
else
attr->compress = true;
if (aoEntry->compresstype != NULL)
attr->compressType = aoEntry->compresstype;
else
attr->compressType = "none";
attr->compressLevel = aoEntry->compresslevel;
attr->checksum = aoEntry->checksum;
attr->safeFSWriteSize = aoEntry->safefswritesize;
attr->splitsize = aoEntry->splitsize;
attr->version = aoEntry->version;
AORelationVersion_CheckValid(attr->version);
scan->proj = proj;
scan->pqs_tupDesc = (relationTupleDesc == NULL) ? RelationGetDescr(relation) : relationTupleDesc;
scan->hawqAttrToParquetColChunks = (int*)palloc0(scan->pqs_tupDesc->natts * sizeof(int));
initscan(scan);
return scan ;
}
/*
* This function initializes a part and returns true if a new index has been prepared for scanning.
*/
static bool
initNextIndexToScan(DynamicIndexScanState *node)
{
IndexScanState *indexState = &(node->indexScanState);
DynamicIndexScan *dynamicIndexScan = (DynamicIndexScan *)node->indexScanState.ss.ps.plan;
EState *estate = indexState->ss.ps.state;
/* Load new index when the scanning of the previous index is done. */
if (indexState->ss.scan_state == SCAN_INIT ||
indexState->ss.scan_state == SCAN_DONE)
{
/* This is the oid of a partition of the table (*not* index) */
Oid *pid = hash_seq_search(&node->pidxStatus);
if (pid == NULL)
{
/* Return if all parts have been scanned. */
node->shouldCallHashSeqTerm = false;
return false;
}
/* Collect number of partitions scanned in EXPLAIN ANALYZE */
if(NULL != indexState->ss.ps.instrument)
{
Instrumentation *instr = indexState->ss.ps.instrument;
instr->numPartScanned ++;
}
DynamicIndexScan_ReMapColumns(node, *pid);
/*
* The is the oid of the partition of an *index*. Note: a partitioned table
* has a root and a set of partitions (may be multi-level). An index
* on a partitioned table also has a root and a set of index partitions.
* We started at table level, and now we are fetching the oid of an index
* partition.
*/
Relation currentRelation = OpenScanRelationByOid(*pid);
indexState->ss.ss_currentRelation = currentRelation;
indexState->ss.ss_ScanTupleSlot->tts_tableOid = *pid;
ExecAssignScanType(&indexState->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&indexState->ss.ps);
ExecAssignScanProjectionInfo(&indexState->ss);
MemoryContextReset(node->partitionMemoryContext);
MemoryContext oldCxt = MemoryContextSwitchTo(node->partitionMemoryContext);
/* Initialize child expressions */
indexState->ss.ps.qual = (List *) ExecInitExpr((Expr *) indexState->ss.ps.plan->qual, (PlanState *) indexState);
indexState->ss.ps.targetlist = (List *) ExecInitExpr((Expr *) indexState->ss.ps.plan->targetlist, (PlanState *) indexState);
Oid pindex = getPhysicalIndexRelid(currentRelation, dynamicIndexScan->logicalIndexInfo);
Assert(OidIsValid(pindex));
indexState->iss_RelationDesc =
OpenIndexRelation(estate, pindex, *pid);
/*
* build the index scan keys from the index qualification
*/
ExecIndexBuildScanKeys((PlanState *) indexState,
indexState->iss_RelationDesc,
dynamicIndexScan->indexqual,
dynamicIndexScan->indexstrategy,
dynamicIndexScan->indexsubtype,
&indexState->iss_ScanKeys,
&indexState->iss_NumScanKeys,
&indexState->iss_RuntimeKeys,
&indexState->iss_NumRuntimeKeys,
NULL,
NULL);
MemoryContextSwitchTo(oldCxt);
if (indexState->iss_NumRuntimeKeys != 0)
{
ExecIndexEvalRuntimeKeys(indexState->iss_RuntimeContext,
indexState->iss_RuntimeKeys,
indexState->iss_NumRuntimeKeys);
}
indexState->iss_RuntimeKeysReady = true;
indexState->iss_ScanDesc = index_beginscan(currentRelation,
indexState->iss_RelationDesc,
estate->es_snapshot,
indexState->iss_NumScanKeys,
indexState->iss_ScanKeys);
indexState->ss.scan_state = SCAN_SCAN;
}
return true;
}
/*
* OperatorCreate
*
* "X" indicates an optional argument (i.e. one that can be NULL or 0)
* operatorName name for new operator
* operatorNamespace namespace for new operator
* leftTypeId X left type ID
* rightTypeId X right type ID
* procedureId procedure ID for operator
* commutatorName X commutator operator
* negatorName X negator operator
* restrictionId X restriction selectivity procedure ID
* joinId X join selectivity procedure ID
* canMerge merge join can be used with this operator
* canHash hash join can be used with this operator
*
* The caller should have validated properties and permissions for the
* objects passed as OID references. We must handle the commutator and
* negator operator references specially, however, since those need not
* exist beforehand.
*
* This routine gets complicated because it allows the user to
* specify operators that do not exist. For example, if operator
* "op" is being defined, the negator operator "negop" and the
* commutator "commop" can also be defined without specifying
* any information other than their names. Since in order to
* add "op" to the PG_OPERATOR catalog, all the Oid's for these
* operators must be placed in the fields of "op", a forward
* declaration is done on the commutator and negator operators.
* This is called creating a shell, and its main effect is to
* create a tuple in the PG_OPERATOR catalog with minimal
* information about the operator (just its name and types).
* Forward declaration is used only for this purpose, it is
* not available to the user as it is for type definition.
*/
ObjectAddress
OperatorCreate(const char *operatorName,
Oid operatorNamespace,
Oid leftTypeId,
Oid rightTypeId,
Oid procedureId,
List *commutatorName,
List *negatorName,
Oid restrictionId,
Oid joinId,
bool canMerge,
bool canHash)
{
Relation pg_operator_desc;
HeapTuple tup;
bool isUpdate;
bool nulls[Natts_pg_operator];
bool replaces[Natts_pg_operator];
Datum values[Natts_pg_operator];
Oid operatorObjectId;
bool operatorAlreadyDefined;
Oid operResultType;
Oid commutatorId,
negatorId;
bool selfCommutator = false;
NameData oname;
int i;
ObjectAddress address;
/*
* Sanity checks
*/
if (!validOperatorName(operatorName))
ereport(ERROR,
(errcode(ERRCODE_INVALID_NAME),
errmsg("\"%s\" is not a valid operator name",
operatorName)));
if (!(OidIsValid(leftTypeId) && OidIsValid(rightTypeId)))
{
/* If it's not a binary op, these things mustn't be set: */
if (commutatorName)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only binary operators can have commutators")));
if (OidIsValid(joinId))
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only binary operators can have join selectivity")));
if (canMerge)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only binary operators can merge join")));
if (canHash)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only binary operators can hash")));
}
operResultType = get_func_rettype(procedureId);
if (operResultType != BOOLOID)
{
/* If it's not a boolean op, these things mustn't be set: */
if (negatorName)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only boolean operators can have negators")));
if (OidIsValid(restrictionId))
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only boolean operators can have restriction selectivity")));
if (OidIsValid(joinId))
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only boolean operators can have join selectivity")));
if (canMerge)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only boolean operators can merge join")));
if (canHash)
ereport(ERROR,
(errcode(ERRCODE_INVALID_FUNCTION_DEFINITION),
errmsg("only boolean operators can hash")));
}
operatorObjectId = OperatorGet(operatorName,
operatorNamespace,
leftTypeId,
rightTypeId,
&operatorAlreadyDefined);
if (operatorAlreadyDefined)
ereport(ERROR,
(errcode(ERRCODE_DUPLICATE_FUNCTION),
errmsg("operator %s already exists",
operatorName)));
/*
* At this point, if operatorObjectId is not InvalidOid then we are
* filling in a previously-created shell. Insist that the user own any
* such shell.
*/
if (OidIsValid(operatorObjectId) &&
!pg_oper_ownercheck(operatorObjectId, GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, OBJECT_OPERATOR,
operatorName);
/*
* Set up the other operators. If they do not currently exist, create
* shells in order to get ObjectId's.
*/
if (commutatorName)
{
/* commutator has reversed arg types */
commutatorId = get_other_operator(commutatorName,
rightTypeId, leftTypeId,
operatorName, operatorNamespace,
leftTypeId, rightTypeId,
true);
/* Permission check: must own other operator */
if (OidIsValid(commutatorId) &&
!pg_oper_ownercheck(commutatorId, GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, OBJECT_OPERATOR,
NameListToString(commutatorName));
/*
* self-linkage to this operator; will fix below. Note that only
* self-linkage for commutation makes sense.
*/
if (!OidIsValid(commutatorId))
selfCommutator = true;
}
else
commutatorId = InvalidOid;
if (negatorName)
{
/* negator has same arg types */
negatorId = get_other_operator(negatorName,
leftTypeId, rightTypeId,
operatorName, operatorNamespace,
leftTypeId, rightTypeId,
false);
/* Permission check: must own other operator */
if (OidIsValid(negatorId) &&
!pg_oper_ownercheck(negatorId, GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, OBJECT_OPERATOR,
NameListToString(negatorName));
}
else
negatorId = InvalidOid;
/*
* set up values in the operator tuple
*/
for (i = 0; i < Natts_pg_operator; ++i)
{
values[i] = (Datum) NULL;
replaces[i] = true;
nulls[i] = false;
}
namestrcpy(&oname, operatorName);
values[Anum_pg_operator_oprname - 1] = NameGetDatum(&oname);
values[Anum_pg_operator_oprnamespace - 1] = ObjectIdGetDatum(operatorNamespace);
values[Anum_pg_operator_oprowner - 1] = ObjectIdGetDatum(GetUserId());
values[Anum_pg_operator_oprkind - 1] = CharGetDatum(leftTypeId ? (rightTypeId ? 'b' : 'r') : 'l');
values[Anum_pg_operator_oprcanmerge - 1] = BoolGetDatum(canMerge);
values[Anum_pg_operator_oprcanhash - 1] = BoolGetDatum(canHash);
values[Anum_pg_operator_oprleft - 1] = ObjectIdGetDatum(leftTypeId);
values[Anum_pg_operator_oprright - 1] = ObjectIdGetDatum(rightTypeId);
values[Anum_pg_operator_oprresult - 1] = ObjectIdGetDatum(operResultType);
values[Anum_pg_operator_oprcom - 1] = ObjectIdGetDatum(commutatorId);
values[Anum_pg_operator_oprnegate - 1] = ObjectIdGetDatum(negatorId);
values[Anum_pg_operator_oprcode - 1] = ObjectIdGetDatum(procedureId);
values[Anum_pg_operator_oprrest - 1] = ObjectIdGetDatum(restrictionId);
values[Anum_pg_operator_oprjoin - 1] = ObjectIdGetDatum(joinId);
pg_operator_desc = heap_open(OperatorRelationId, RowExclusiveLock);
/*
* If we are replacing an operator shell, update; else insert
*/
if (operatorObjectId)
{
isUpdate = true;
tup = SearchSysCacheCopy1(OPEROID,
ObjectIdGetDatum(operatorObjectId));
if (!HeapTupleIsValid(tup))
elog(ERROR, "cache lookup failed for operator %u",
operatorObjectId);
replaces[Anum_pg_operator_oid - 1] = false;
tup = heap_modify_tuple(tup,
RelationGetDescr(pg_operator_desc),
values,
nulls,
replaces);
CatalogTupleUpdate(pg_operator_desc, &tup->t_self, tup);
}
else
{
isUpdate = false;
operatorObjectId = GetNewOidWithIndex(pg_operator_desc,
OperatorOidIndexId,
Anum_pg_operator_oid);
values[Anum_pg_operator_oid - 1] = ObjectIdGetDatum(operatorObjectId);
tup = heap_form_tuple(RelationGetDescr(pg_operator_desc),
values, nulls);
CatalogTupleInsert(pg_operator_desc, tup);
}
/* Add dependencies for the entry */
address = makeOperatorDependencies(tup, isUpdate);
/* Post creation hook for new operator */
InvokeObjectPostCreateHook(OperatorRelationId, operatorObjectId, 0);
heap_close(pg_operator_desc, RowExclusiveLock);
/*
* If a commutator and/or negator link is provided, update the other
* operator(s) to point at this one, if they don't already have a link.
* This supports an alternative style of operator definition wherein the
* user first defines one operator without giving negator or commutator,
* then defines the other operator of the pair with the proper commutator
* or negator attribute. That style doesn't require creation of a shell,
* and it's the only style that worked right before Postgres version 6.5.
* This code also takes care of the situation where the new operator is
* its own commutator.
*/
if (selfCommutator)
commutatorId = operatorObjectId;
if (OidIsValid(commutatorId) || OidIsValid(negatorId))
OperatorUpd(operatorObjectId, commutatorId, negatorId, false);
return address;
}
/*
* Change tablespace owner
*/
void
AlterTableSpaceOwner(const char *name, Oid newOwnerId)
{
Relation rel;
ScanKeyData entry[1];
HeapScanDesc scandesc;
Form_pg_tablespace spcForm;
HeapTuple tup;
/* Search pg_tablespace */
rel = heap_open(TableSpaceRelationId, RowExclusiveLock);
ScanKeyInit(&entry[0],
Anum_pg_tablespace_spcname,
BTEqualStrategyNumber, F_NAMEEQ,
CStringGetDatum(name));
scandesc = heap_beginscan(rel, SnapshotNow, 1, entry);
tup = heap_getnext(scandesc, ForwardScanDirection);
if (!HeapTupleIsValid(tup))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("tablespace \"%s\" does not exist", name)));
spcForm = (Form_pg_tablespace) GETSTRUCT(tup);
/*
* If the new owner is the same as the existing owner, consider the
* command to have succeeded. This is for dump restoration purposes.
*/
if (spcForm->spcowner != newOwnerId)
{
Datum repl_val[Natts_pg_tablespace];
bool repl_null[Natts_pg_tablespace];
bool repl_repl[Natts_pg_tablespace];
Acl *newAcl;
Datum aclDatum;
bool isNull;
HeapTuple newtuple;
/* Otherwise, must be owner of the existing object */
if (!pg_tablespace_ownercheck(HeapTupleGetOid(tup), GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_TABLESPACE,
name);
/* Must be able to become new owner */
check_is_member_of_role(GetUserId(), newOwnerId);
/*
* Normally we would also check for create permissions here, but there
* are none for tablespaces so we follow what rename tablespace does
* and omit the create permissions check.
*
* NOTE: Only superusers may create tablespaces to begin with and so
* initially only a superuser would be able to change its ownership
* anyway.
*/
memset(repl_null, false, sizeof(repl_null));
memset(repl_repl, false, sizeof(repl_repl));
repl_repl[Anum_pg_tablespace_spcowner - 1] = true;
repl_val[Anum_pg_tablespace_spcowner - 1] = ObjectIdGetDatum(newOwnerId);
/*
* Determine the modified ACL for the new owner. This is only
* necessary when the ACL is non-null.
*/
aclDatum = heap_getattr(tup,
Anum_pg_tablespace_spcacl,
RelationGetDescr(rel),
&isNull);
if (!isNull)
{
newAcl = aclnewowner(DatumGetAclP(aclDatum),
spcForm->spcowner, newOwnerId);
repl_repl[Anum_pg_tablespace_spcacl - 1] = true;
repl_val[Anum_pg_tablespace_spcacl - 1] = PointerGetDatum(newAcl);
}
newtuple = heap_modify_tuple(tup, RelationGetDescr(rel), repl_val, repl_null, repl_repl);
simple_heap_update(rel, &newtuple->t_self, newtuple);
CatalogUpdateIndexes(rel, newtuple);
heap_freetuple(newtuple);
/* Update owner dependency reference */
changeDependencyOnOwner(TableSpaceRelationId, HeapTupleGetOid(tup),
newOwnerId);
}
heap_endscan(scandesc);
heap_close(rel, NoLock);
}
/*
* initNextTableToScan
* Find the next table to scan and initiate the scan if the previous table
* is finished.
*
* If scanning on the current table is not finished, or a new table is found,
* this function returns true.
* If no more table is found, this function returns false.
*/
static bool
initNextTableToScan(DynamicTableScanState *node)
{
ScanState *scanState = (ScanState *)node;
if (scanState->scan_state == SCAN_INIT ||
scanState->scan_state == SCAN_DONE)
{
Oid *pid = hash_seq_search(&node->pidStatus);
if (pid == NULL)
{
node->shouldCallHashSeqTerm = false;
return false;
}
/* Collect number of partitions scanned in EXPLAIN ANALYZE */
if (NULL != scanState->ps.instrument)
{
Instrumentation *instr = scanState->ps.instrument;
instr->numPartScanned ++;
}
/*
* Inside ExecInitScanTupleSlot() we set the tuple table slot's oid
* to range table entry's relid, which for partitioned table always set
* to parent table's oid. In queries where we need to read table oids
* (MPP-20736) we use the tuple table slot's saved oid (refer to slot_getsysattr()).
* This wrongly returns parent oid, instead of partition oid. Therefore,
* to return correct partition oid, we need to update
* our tuple table slot's oid to reflect the partition oid.
*/
scanState->ss_ScanTupleSlot->tts_tableOid = *pid;
scanState->ss_currentRelation = OpenScanRelationByOid(*pid);
Relation lastScannedRel = OpenScanRelationByOid(node->lastRelOid);
TupleDesc lastTupDesc = RelationGetDescr(lastScannedRel);
CloseScanRelation(lastScannedRel);
TupleDesc partTupDesc = RelationGetDescr(scanState->ss_currentRelation);
ExecAssignScanType(scanState, partTupDesc);
AttrNumber *attMap = NULL;
attMap = varattnos_map(lastTupDesc, partTupDesc);
/* If attribute remapping is not necessary, then do not change the varattno */
if (attMap)
{
change_varattnos_of_a_varno((Node*)scanState->ps.plan->qual, attMap, node->scanrelid);
change_varattnos_of_a_varno((Node*)scanState->ps.plan->targetlist, attMap, node->scanrelid);
/*
* Now that the varattno mapping has been changed, change the relation that
* the new varnos correspond to
*/
node->lastRelOid = *pid;
}
/*
* For the very first partition, the targetlist of planstate is set to null. So, we must
* initialize quals and targetlist, regardless of remapping requirements. For later
* partitions, we only initialize quals and targetlist if a column re-mapping is necessary.
*/
if (attMap || node->firstPartition)
{
node->firstPartition = false;
MemoryContextReset(node->partitionMemoryContext);
MemoryContext oldCxt = MemoryContextSwitchTo(node->partitionMemoryContext);
/* Initialize child expressions */
scanState->ps.qual = (List *)ExecInitExpr((Expr *)scanState->ps.plan->qual, (PlanState*)scanState);
scanState->ps.targetlist = (List *)ExecInitExpr((Expr *)scanState->ps.plan->targetlist, (PlanState*)scanState);
MemoryContextSwitchTo(oldCxt);
}
if (attMap)
{
pfree(attMap);
}
ExecAssignScanProjectionInfo(scanState);
scanState->tableType = getTableType(scanState->ss_currentRelation);
BeginTableScanRelation(scanState);
}
return true;
}
/*
* Extract all item values from a BRIN index page
*
* Usage: SELECT * FROM brin_page_items(get_raw_page('idx', 1), 'idx'::regclass);
*/
Datum
brin_page_items(PG_FUNCTION_ARGS)
{
brin_page_state *state;
FuncCallContext *fctx;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use raw page functions"))));
if (SRF_IS_FIRSTCALL())
{
bytea *raw_page = PG_GETARG_BYTEA_P(0);
Oid indexRelid = PG_GETARG_OID(1);
Page page;
TupleDesc tupdesc;
MemoryContext mctx;
Relation indexRel;
AttrNumber attno;
/* minimally verify the page we got */
page = verify_brin_page(raw_page, BRIN_PAGETYPE_REGULAR, "regular");
/* create a function context for cross-call persistence */
fctx = SRF_FIRSTCALL_INIT();
/* switch to memory context appropriate for multiple function calls */
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
indexRel = index_open(indexRelid, AccessShareLock);
state = palloc(offsetof(brin_page_state, columns) +
sizeof(brin_column_state) * RelationGetDescr(indexRel)->natts);
state->bdesc = brin_build_desc(indexRel);
state->page = page;
state->offset = FirstOffsetNumber;
state->unusedItem = false;
state->done = false;
state->dtup = NULL;
/*
* Initialize output functions for all indexed datatypes; simplifies
* calling them later.
*/
for (attno = 1; attno <= state->bdesc->bd_tupdesc->natts; attno++)
{
Oid output;
bool isVarlena;
BrinOpcInfo *opcinfo;
int i;
brin_column_state *column;
opcinfo = state->bdesc->bd_info[attno - 1];
column = palloc(offsetof(brin_column_state, outputFn) +
sizeof(FmgrInfo) * opcinfo->oi_nstored);
column->nstored = opcinfo->oi_nstored;
for (i = 0; i < opcinfo->oi_nstored; i++)
{
getTypeOutputInfo(opcinfo->oi_typcache[i]->type_id, &output, &isVarlena);
fmgr_info(output, &column->outputFn[i]);
}
state->columns[attno - 1] = column;
}
index_close(indexRel, AccessShareLock);
fctx->user_fctx = state;
fctx->tuple_desc = BlessTupleDesc(tupdesc);
MemoryContextSwitchTo(mctx);
}
fctx = SRF_PERCALL_SETUP();
state = fctx->user_fctx;
if (!state->done)
{
HeapTuple result;
Datum values[7];
bool nulls[7];
/*
* This loop is called once for every attribute of every tuple in the
* page. At the start of a tuple, we get a NULL dtup; that's our
* signal for obtaining and decoding the next one. If that's not the
* case, we output the next attribute.
*/
if (state->dtup == NULL)
{
BrinTuple *tup;
MemoryContext mctx;
ItemId itemId;
/* deformed tuple must live across calls */
mctx = MemoryContextSwitchTo(fctx->multi_call_memory_ctx);
/* verify item status: if there's no data, we can't decode */
itemId = PageGetItemId(state->page, state->offset);
if (ItemIdIsUsed(itemId))
{
tup = (BrinTuple *) PageGetItem(state->page,
PageGetItemId(state->page,
state->offset));
state->dtup = brin_deform_tuple(state->bdesc, tup);
state->attno = 1;
state->unusedItem = false;
}
else
state->unusedItem = true;
MemoryContextSwitchTo(mctx);
}
else
state->attno++;
MemSet(nulls, 0, sizeof(nulls));
if (state->unusedItem)
{
values[0] = UInt16GetDatum(state->offset);
nulls[1] = true;
nulls[2] = true;
nulls[3] = true;
nulls[4] = true;
nulls[5] = true;
nulls[6] = true;
}
else
{
int att = state->attno - 1;
values[0] = UInt16GetDatum(state->offset);
values[1] = UInt32GetDatum(state->dtup->bt_blkno);
values[2] = UInt16GetDatum(state->attno);
values[3] = BoolGetDatum(state->dtup->bt_columns[att].bv_allnulls);
values[4] = BoolGetDatum(state->dtup->bt_columns[att].bv_hasnulls);
values[5] = BoolGetDatum(state->dtup->bt_placeholder);
if (!state->dtup->bt_columns[att].bv_allnulls)
{
BrinValues *bvalues = &state->dtup->bt_columns[att];
StringInfoData s;
bool first;
int i;
initStringInfo(&s);
appendStringInfoChar(&s, '{');
first = true;
for (i = 0; i < state->columns[att]->nstored; i++)
{
char *val;
if (!first)
appendStringInfoString(&s, " .. ");
first = false;
val = OutputFunctionCall(&state->columns[att]->outputFn[i],
bvalues->bv_values[i]);
appendStringInfoString(&s, val);
pfree(val);
}
appendStringInfoChar(&s, '}');
values[6] = CStringGetTextDatum(s.data);
pfree(s.data);
}
else
{
nulls[6] = true;
}
}
result = heap_form_tuple(fctx->tuple_desc, values, nulls);
/*
* If the item was unused, jump straight to the next one; otherwise,
* the only cleanup needed here is to set our signal to go to the next
* tuple in the following iteration, by freeing the current one.
*/
if (state->unusedItem)
state->offset = OffsetNumberNext(state->offset);
else if (state->attno >= state->bdesc->bd_tupdesc->natts)
{
pfree(state->dtup);
state->dtup = NULL;
state->offset = OffsetNumberNext(state->offset);
}
/*
* If we're beyond the end of the page, set flag to end the function
* in the following iteration.
*/
if (state->offset > PageGetMaxOffsetNumber(state->page))
state->done = true;
SRF_RETURN_NEXT(fctx, HeapTupleGetDatum(result));
}
brin_free_desc(state->bdesc);
SRF_RETURN_DONE(fctx);
}
/*
* Extract all item values from a BRIN index page
*
* Usage: SELECT * FROM brin_page_items(get_raw_page('idx', 1), 'idx'::regclass);
*/
Datum
brin_page_items(PG_FUNCTION_ARGS)
{
bytea *raw_page = PG_GETARG_BYTEA_P(0);
Oid indexRelid = PG_GETARG_OID(1);
ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo;
TupleDesc tupdesc;
MemoryContext oldcontext;
Tuplestorestate *tupstore;
Relation indexRel;
brin_column_state **columns;
BrinDesc *bdesc;
BrinMemTuple *dtup;
Page page;
OffsetNumber offset;
AttrNumber attno;
bool unusedItem;
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to use raw page functions"))));
/* check to see if caller supports us returning a tuplestore */
if (rsinfo == NULL || !IsA(rsinfo, ReturnSetInfo))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("set-valued function called in context that cannot accept a set")));
if (!(rsinfo->allowedModes & SFRM_Materialize) ||
rsinfo->expectedDesc == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("materialize mode required, but it is not allowed in this context")));
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
/* Build tuplestore to hold the result rows */
oldcontext = MemoryContextSwitchTo(rsinfo->econtext->ecxt_per_query_memory);
tupstore = tuplestore_begin_heap(true, false, work_mem);
rsinfo->returnMode = SFRM_Materialize;
rsinfo->setResult = tupstore;
rsinfo->setDesc = tupdesc;
MemoryContextSwitchTo(oldcontext);
indexRel = index_open(indexRelid, AccessShareLock);
bdesc = brin_build_desc(indexRel);
/* minimally verify the page we got */
page = verify_brin_page(raw_page, BRIN_PAGETYPE_REGULAR, "regular");
/*
* Initialize output functions for all indexed datatypes; simplifies
* calling them later.
*/
columns = palloc(sizeof(brin_column_state *) * RelationGetDescr(indexRel)->natts);
for (attno = 1; attno <= bdesc->bd_tupdesc->natts; attno++)
{
Oid output;
bool isVarlena;
BrinOpcInfo *opcinfo;
int i;
brin_column_state *column;
opcinfo = bdesc->bd_info[attno - 1];
column = palloc(offsetof(brin_column_state, outputFn) +
sizeof(FmgrInfo) * opcinfo->oi_nstored);
column->nstored = opcinfo->oi_nstored;
for (i = 0; i < opcinfo->oi_nstored; i++)
{
getTypeOutputInfo(opcinfo->oi_typcache[i]->type_id, &output, &isVarlena);
fmgr_info(output, &column->outputFn[i]);
}
columns[attno - 1] = column;
}
offset = FirstOffsetNumber;
unusedItem = false;
dtup = NULL;
for (;;)
{
Datum values[7];
bool nulls[7];
/*
* This loop is called once for every attribute of every tuple in the
* page. At the start of a tuple, we get a NULL dtup; that's our
* signal for obtaining and decoding the next one. If that's not the
* case, we output the next attribute.
*/
if (dtup == NULL)
{
ItemId itemId;
/* verify item status: if there's no data, we can't decode */
itemId = PageGetItemId(page, offset);
if (ItemIdIsUsed(itemId))
{
dtup = brin_deform_tuple(bdesc,
(BrinTuple *) PageGetItem(page, itemId),
NULL);
attno = 1;
unusedItem = false;
}
else
unusedItem = true;
}
else
attno++;
MemSet(nulls, 0, sizeof(nulls));
if (unusedItem)
{
values[0] = UInt16GetDatum(offset);
nulls[1] = true;
nulls[2] = true;
nulls[3] = true;
nulls[4] = true;
nulls[5] = true;
nulls[6] = true;
}
else
{
int att = attno - 1;
values[0] = UInt16GetDatum(offset);
values[1] = UInt32GetDatum(dtup->bt_blkno);
values[2] = UInt16GetDatum(attno);
values[3] = BoolGetDatum(dtup->bt_columns[att].bv_allnulls);
values[4] = BoolGetDatum(dtup->bt_columns[att].bv_hasnulls);
values[5] = BoolGetDatum(dtup->bt_placeholder);
if (!dtup->bt_columns[att].bv_allnulls)
{
BrinValues *bvalues = &dtup->bt_columns[att];
StringInfoData s;
bool first;
int i;
initStringInfo(&s);
appendStringInfoChar(&s, '{');
first = true;
for (i = 0; i < columns[att]->nstored; i++)
{
char *val;
if (!first)
appendStringInfoString(&s, " .. ");
first = false;
val = OutputFunctionCall(&columns[att]->outputFn[i],
bvalues->bv_values[i]);
appendStringInfoString(&s, val);
pfree(val);
}
appendStringInfoChar(&s, '}');
values[6] = CStringGetTextDatum(s.data);
pfree(s.data);
}
else
{
nulls[6] = true;
}
}
tuplestore_putvalues(tupstore, tupdesc, values, nulls);
/*
* If the item was unused, jump straight to the next one; otherwise,
* the only cleanup needed here is to set our signal to go to the next
* tuple in the following iteration, by freeing the current one.
*/
if (unusedItem)
offset = OffsetNumberNext(offset);
else if (attno >= bdesc->bd_tupdesc->natts)
{
pfree(dtup);
dtup = NULL;
offset = OffsetNumberNext(offset);
}
/*
* If we're beyond the end of the page, we're done.
*/
if (offset > PageGetMaxOffsetNumber(page))
break;
}
/* clean up and return the tuplestore */
brin_free_desc(bdesc);
tuplestore_donestoring(tupstore);
index_close(indexRel, AccessShareLock);
return (Datum) 0;
}
/*
* Guts of language creation.
*/
static ObjectAddress
create_proc_lang(const char *languageName, bool replace,
Oid languageOwner, Oid handlerOid, Oid inlineOid,
Oid valOid, bool trusted)
{
Relation rel;
TupleDesc tupDesc;
Datum values[Natts_pg_language];
bool nulls[Natts_pg_language];
bool replaces[Natts_pg_language];
NameData langname;
HeapTuple oldtup;
HeapTuple tup;
bool is_update;
ObjectAddress myself,
referenced;
rel = heap_open(LanguageRelationId, RowExclusiveLock);
tupDesc = RelationGetDescr(rel);
/* Prepare data to be inserted */
memset(values, 0, sizeof(values));
memset(nulls, false, sizeof(nulls));
memset(replaces, true, sizeof(replaces));
namestrcpy(&langname, languageName);
values[Anum_pg_language_lanname - 1] = NameGetDatum(&langname);
values[Anum_pg_language_lanowner - 1] = ObjectIdGetDatum(languageOwner);
values[Anum_pg_language_lanispl - 1] = BoolGetDatum(true);
values[Anum_pg_language_lanpltrusted - 1] = BoolGetDatum(trusted);
values[Anum_pg_language_lanplcallfoid - 1] = ObjectIdGetDatum(handlerOid);
values[Anum_pg_language_laninline - 1] = ObjectIdGetDatum(inlineOid);
values[Anum_pg_language_lanvalidator - 1] = ObjectIdGetDatum(valOid);
nulls[Anum_pg_language_lanacl - 1] = true;
/* Check for pre-existing definition */
oldtup = SearchSysCache1(LANGNAME, PointerGetDatum(languageName));
if (HeapTupleIsValid(oldtup))
{
/* There is one; okay to replace it? */
if (!replace)
ereport(ERROR,
(errcode(ERRCODE_DUPLICATE_OBJECT),
errmsg("language \"%s\" already exists", languageName)));
if (!pg_language_ownercheck(HeapTupleGetOid(oldtup), languageOwner))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_LANGUAGE,
languageName);
/*
* Do not change existing ownership or permissions. Note
* dependency-update code below has to agree with this decision.
*/
replaces[Anum_pg_language_lanowner - 1] = false;
replaces[Anum_pg_language_lanacl - 1] = false;
/* Okay, do it... */
tup = heap_modify_tuple(oldtup, tupDesc, values, nulls, replaces);
simple_heap_update(rel, &tup->t_self, tup);
ReleaseSysCache(oldtup);
is_update = true;
}
else
{
/* Creating a new language */
tup = heap_form_tuple(tupDesc, values, nulls);
simple_heap_insert(rel, tup);
is_update = false;
}
/* Need to update indexes for either the insert or update case */
CatalogUpdateIndexes(rel, tup);
/*
* Create dependencies for the new language. If we are updating an
* existing language, first delete any existing pg_depend entries.
* (However, since we are not changing ownership or permissions, the
* shared dependencies do *not* need to change, and we leave them alone.)
*/
myself.classId = LanguageRelationId;
myself.objectId = HeapTupleGetOid(tup);
myself.objectSubId = 0;
if (is_update)
deleteDependencyRecordsFor(myself.classId, myself.objectId, true);
/* dependency on owner of language */
if (!is_update)
recordDependencyOnOwner(myself.classId, myself.objectId,
languageOwner);
/* dependency on extension */
recordDependencyOnCurrentExtension(&myself, is_update);
/* dependency on the PL handler function */
referenced.classId = ProcedureRelationId;
referenced.objectId = handlerOid;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
/* dependency on the inline handler function, if any */
if (OidIsValid(inlineOid))
{
referenced.classId = ProcedureRelationId;
referenced.objectId = inlineOid;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
}
/* dependency on the validator function, if any */
if (OidIsValid(valOid))
{
referenced.classId = ProcedureRelationId;
referenced.objectId = valOid;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
}
/* Post creation hook for new procedural language */
InvokeObjectPostCreateHook(LanguageRelationId, myself.objectId, 0);
heap_close(rel, RowExclusiveLock);
return myself;
}
/* ----------------------------------------------------------------
* ExecInitIndexScan
*
* Initializes the index scan's state information, creates
* scan keys, and opens the base and index relations.
*
* Note: index scans have 2 sets of state information because
* we have to keep track of the base relation and the
* index relation.
* ----------------------------------------------------------------
*/
IndexScanState *
ExecInitIndexScan(IndexScan *node, EState *estate, int eflags)
{
IndexScanState *indexstate;
Relation currentRelation;
bool relistarget;
/*
* create state structure
*/
indexstate = makeNode(IndexScanState);
indexstate->ss.ps.plan = (Plan *) node;
indexstate->ss.ps.state = estate;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &indexstate->ss.ps);
/*indexstate->ss.ps.ps_TupFromTlist = false;*/
/*
* initialize child expressions
*
* Note: we don't initialize all of the indexqual expression, only the
* sub-parts corresponding to runtime keys (see below). The indexqualorig
* expression is always initialized even though it will only be used in
* some uncommon cases --- would be nice to improve that. (Problem is
* that any SubPlans present in the expression must be found now...)
*/
indexstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) indexstate);
indexstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) indexstate);
indexstate->indexqualorig = (List *)
ExecInitExpr((Expr *) node->indexqualorig,
(PlanState *) indexstate);
#define INDEXSCAN_NSLOTS 2
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &indexstate->ss.ps);
ExecInitScanTupleSlot(estate, &indexstate->ss);
/*
* open the base relation and acquire appropriate lock on it.
*/
currentRelation = ExecOpenScanRelation(estate, node->scan.scanrelid);
indexstate->ss.ss_currentRelation = currentRelation;
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&indexstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&indexstate->ss.ps);
ExecAssignScanProjectionInfo(&indexstate->ss);
/*
* If we are just doing EXPLAIN (ie, aren't going to run the plan), stop
* here. This allows an index-advisor plugin to EXPLAIN a plan containing
* references to nonexistent indexes.
*/
if (eflags & EXEC_FLAG_EXPLAIN_ONLY)
return indexstate;
/*
* Open the index relation.
*
* If the parent table is one of the target relations of the query, then
* InitPlan already opened and write-locked the index, so we can avoid
* taking another lock here. Otherwise we need a normal reader's lock.
*/
relistarget = ExecRelationIsTargetRelation(estate, node->scan.scanrelid);
indexstate->iss_RelationDesc = index_open(node->indexid,
relistarget ? NoLock : AccessShareLock);
/*
* build the index scan keys from the index qualification
*/
ExecIndexBuildScanKeys((PlanState *) indexstate,
indexstate->iss_RelationDesc,
node->indexqual,
node->indexstrategy,
node->indexsubtype,
&indexstate->iss_ScanKeys,
&indexstate->iss_NumScanKeys,
&indexstate->iss_RuntimeKeys,
&indexstate->iss_NumRuntimeKeys,
NULL, /* no ArrayKeys */
NULL);
InitRuntimeKeysContext(indexstate);
Assert(NULL != indexstate->iss_RuntimeContext);
/*
* Initialize index-specific scan state
*/
indexstate->iss_RuntimeKeysReady = false;
initGpmonPktForIndexScan((Plan *)node, &indexstate->ss.ps.gpmon_pkt, estate);
/*
* If eflag contains EXEC_FLAG_REWIND or EXEC_FLAG_BACKWARD or EXEC_FLAG_MARK,
* then this node is not eager free safe.
*/
indexstate->ss.ps.delayEagerFree =
((eflags & (EXEC_FLAG_REWIND | EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)) != 0);
/*
* all done.
*/
return indexstate;
}
/*
* RangeCreate
* Create an entry in pg_range.
*/
void
RangeCreate(Oid rangeTypeOid, Oid rangeSubType, Oid rangeCollation,
Oid rangeSubOpclass, RegProcedure rangeCanonical,
RegProcedure rangeSubDiff)
{
Relation pg_range;
Datum values[Natts_pg_range];
bool nulls[Natts_pg_range];
HeapTuple tup;
ObjectAddress myself;
ObjectAddress referenced;
pg_range = heap_open(RangeRelationId, RowExclusiveLock);
memset(nulls, 0, sizeof(nulls));
values[Anum_pg_range_rngtypid - 1] = ObjectIdGetDatum(rangeTypeOid);
values[Anum_pg_range_rngsubtype - 1] = ObjectIdGetDatum(rangeSubType);
values[Anum_pg_range_rngcollation - 1] = ObjectIdGetDatum(rangeCollation);
values[Anum_pg_range_rngsubopc - 1] = ObjectIdGetDatum(rangeSubOpclass);
values[Anum_pg_range_rngcanonical - 1] = ObjectIdGetDatum(rangeCanonical);
values[Anum_pg_range_rngsubdiff - 1] = ObjectIdGetDatum(rangeSubDiff);
tup = heap_form_tuple(RelationGetDescr(pg_range), values, nulls);
simple_heap_insert(pg_range, tup);
CatalogUpdateIndexes(pg_range, tup);
heap_freetuple(tup);
/* record type's dependencies on range-related items */
myself.classId = TypeRelationId;
myself.objectId = rangeTypeOid;
myself.objectSubId = 0;
referenced.classId = TypeRelationId;
referenced.objectId = rangeSubType;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
referenced.classId = OperatorClassRelationId;
referenced.objectId = rangeSubOpclass;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
if (OidIsValid(rangeCollation))
{
referenced.classId = CollationRelationId;
referenced.objectId = rangeCollation;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
}
if (OidIsValid(rangeCanonical))
{
referenced.classId = ProcedureRelationId;
referenced.objectId = rangeCanonical;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
}
if (OidIsValid(rangeSubDiff))
{
referenced.classId = ProcedureRelationId;
referenced.objectId = rangeSubDiff;
referenced.objectSubId = 0;
recordDependencyOn(&myself, &referenced, DEPENDENCY_NORMAL);
}
heap_close(pg_range, RowExclusiveLock);
}
/* ----------------------------------------------------------------
* ExecInitBitmapHeapScan
*
* Initializes the scan's state information.
* ----------------------------------------------------------------
*/
BitmapHeapScanState *
ExecInitBitmapHeapScan(BitmapHeapScan *node, EState *estate, int eflags)
{
BitmapHeapScanState *scanstate;
Relation currentRelation;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* Assert caller didn't ask for an unsafe snapshot --- see comments at
* head of file.
*/
Assert(IsMVCCSnapshot(estate->es_snapshot));
/*
* create state structure
*/
scanstate = makeNode(BitmapHeapScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->tbm = NULL;
scanstate->tbmiterator = NULL;
scanstate->tbmres = NULL;
scanstate->prefetch_iterator = NULL;
scanstate->prefetch_pages = 0;
scanstate->prefetch_target = 0;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
scanstate->ss.ps.ps_TupFromTlist = false;
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
scanstate->bitmapqualorig = (List *)
ExecInitExpr((Expr *) node->bitmapqualorig,
(PlanState *) scanstate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* open the base relation and acquire appropriate lock on it.
*/
currentRelation = ExecOpenScanRelation(estate, node->scan.scanrelid);
scanstate->ss.ss_currentRelation = currentRelation;
/*
* Even though we aren't going to do a conventional seqscan, it is useful
* to create a HeapScanDesc --- most of the fields in it are usable.
*/
scanstate->ss.ss_currentScanDesc = heap_beginscan_bm(currentRelation,
estate->es_snapshot,
0,
NULL);
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
/*
* initialize child nodes
*
* We do this last because the child nodes will open indexscans on our
* relation's indexes, and we want to be sure we have acquired a lock on
* the relation first.
*/
outerPlanState(scanstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* all done.
*/
return scanstate;
}
/*
* CatalogIndexInsert - insert index entries for one catalog tuple
*
* This should be called for each inserted or updated catalog tuple.
*
* This is effectively a cut-down version of ExecInsertIndexTuples.
*/
void
CatalogIndexInsert(CatalogIndexState indstate, HeapTuple heapTuple)
{
int i;
int numIndexes;
RelationPtr relationDescs;
Relation heapRelation;
TupleTableSlot *slot;
IndexInfo **indexInfoArray;
Datum values[INDEX_MAX_KEYS];
bool isnull[INDEX_MAX_KEYS];
/*
* Get information from the state structure. Fall out if nothing to do.
*/
numIndexes = indstate->ri_NumIndices;
if (numIndexes == 0)
return;
relationDescs = indstate->ri_IndexRelationDescs;
indexInfoArray = indstate->ri_IndexRelationInfo;
heapRelation = indstate->ri_RelationDesc;
/* Need a slot to hold the tuple being examined */
slot = MakeSingleTupleTableSlot(RelationGetDescr(heapRelation));
ExecStoreTuple(heapTuple, slot, InvalidBuffer, false);
/*
* for each index, form and insert the index tuple
*/
for (i = 0; i < numIndexes; i++)
{
IndexInfo *indexInfo;
indexInfo = indexInfoArray[i];
/*
* Expressional and partial indexes on system catalogs are not
* supported
*/
Assert(indexInfo->ii_Expressions == NIL);
Assert(indexInfo->ii_Predicate == NIL);
/*
* FormIndexDatum fills in its values and isnull parameters with the
* appropriate values for the column(s) of the index.
*/
FormIndexDatum(indexInfo,
slot,
NULL, /* no expression eval to do */
values,
isnull);
/*
* The index AM does the rest.
*/
index_insert(relationDescs[i], /* index relation */
values, /* array of index Datums */
isnull, /* is-null flags */
&(heapTuple->t_self), /* tid of heap tuple */
heapRelation,
relationDescs[i]->rd_index->indisunique);
}
ExecDropSingleTupleTableSlot(slot);
}
/* ----------------------------------------------------------------
* ExecInitBitmapHeapScan
*
* Initializes the scan's state information.
* ----------------------------------------------------------------
*/
BitmapHeapScanState *
ExecInitBitmapHeapScan(BitmapHeapScan *node, EState *estate, int eflags)
{
BitmapHeapScanState *scanstate;
Relation currentRelation;
int io_concurrency;
/* check for unsupported flags */
Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
/*
* Assert caller didn't ask for an unsafe snapshot --- see comments at
* head of file.
*/
Assert(IsMVCCSnapshot(estate->es_snapshot));
/*
* create state structure
*/
scanstate = makeNode(BitmapHeapScanState);
scanstate->ss.ps.plan = (Plan *) node;
scanstate->ss.ps.state = estate;
scanstate->tbm = NULL;
scanstate->tbmiterator = NULL;
scanstate->tbmres = NULL;
scanstate->exact_pages = 0;
scanstate->lossy_pages = 0;
scanstate->prefetch_iterator = NULL;
scanstate->prefetch_pages = 0;
scanstate->prefetch_target = 0;
/* may be updated below */
scanstate->prefetch_maximum = target_prefetch_pages;
/*
* Miscellaneous initialization
*
* create expression context for node
*/
ExecAssignExprContext(estate, &scanstate->ss.ps);
scanstate->ss.ps.ps_TupFromTlist = false;
/*
* initialize child expressions
*/
scanstate->ss.ps.targetlist = (List *)
ExecInitExpr((Expr *) node->scan.plan.targetlist,
(PlanState *) scanstate);
scanstate->ss.ps.qual = (List *)
ExecInitExpr((Expr *) node->scan.plan.qual,
(PlanState *) scanstate);
scanstate->bitmapqualorig = (List *)
ExecInitExpr((Expr *) node->bitmapqualorig,
(PlanState *) scanstate);
/*
* tuple table initialization
*/
ExecInitResultTupleSlot(estate, &scanstate->ss.ps);
ExecInitScanTupleSlot(estate, &scanstate->ss);
/*
* open the base relation and acquire appropriate lock on it.
*/
currentRelation = ExecOpenScanRelation(estate, node->scan.scanrelid, eflags);
/*
* Determine the maximum for prefetch_target. If the tablespace has a
* specific IO concurrency set, use that to compute the corresponding
* maximum value; otherwise, we already initialized to the value computed
* by the GUC machinery.
*/
io_concurrency =
get_tablespace_io_concurrency(currentRelation->rd_rel->reltablespace);
if (io_concurrency != effective_io_concurrency)
{
double maximum;
if (ComputeIoConcurrency(io_concurrency, &maximum))
scanstate->prefetch_maximum = rint(maximum);
}
scanstate->ss.ss_currentRelation = currentRelation;
/*
* Even though we aren't going to do a conventional seqscan, it is useful
* to create a HeapScanDesc --- most of the fields in it are usable.
*/
scanstate->ss.ss_currentScanDesc = heap_beginscan_bm(currentRelation,
estate->es_snapshot,
0,
NULL);
/*
* get the scan type from the relation descriptor.
*/
ExecAssignScanType(&scanstate->ss, RelationGetDescr(currentRelation));
/*
* Initialize result tuple type and projection info.
*/
ExecAssignResultTypeFromTL(&scanstate->ss.ps);
ExecAssignScanProjectionInfo(&scanstate->ss);
/*
* initialize child nodes
*
* We do this last because the child nodes will open indexscans on our
* relation's indexes, and we want to be sure we have acquired a lock on
* the relation first.
*/
outerPlanState(scanstate) = ExecInitNode(outerPlan(node), estate, eflags);
/*
* all done.
*/
return scanstate;
}
static void
CatalogCacheInitializeCache(CatCache *cache)
{
Relation relation;
MemoryContext oldcxt;
TupleDesc tupdesc;
int i;
CatalogCacheInitializeCache_DEBUG1;
relation = heap_open(cache->cc_reloid, AccessShareLock);
/*
* switch to the cache context so our allocations do not vanish at the end
* of a transaction
*/
Assert(CacheMemoryContext != NULL);
oldcxt = MemoryContextSwitchTo(CacheMemoryContext);
/*
* copy the relcache's tuple descriptor to permanent cache storage
*/
tupdesc = CreateTupleDescCopyConstr(RelationGetDescr(relation));
/*
* save the relation's name and relisshared flag, too (cc_relname is used
* only for debugging purposes)
*/
cache->cc_relname = pstrdup(RelationGetRelationName(relation));
cache->cc_relisshared = RelationGetForm(relation)->relisshared;
/*
* return to the caller's memory context and close the rel
*/
MemoryContextSwitchTo(oldcxt);
heap_close(relation, AccessShareLock);
CACHE3_elog(DEBUG2, "CatalogCacheInitializeCache: %s, %d keys",
cache->cc_relname, cache->cc_nkeys);
/*
* initialize cache's key information
*/
for (i = 0; i < cache->cc_nkeys; ++i)
{
Oid keytype;
RegProcedure eqfunc;
CatalogCacheInitializeCache_DEBUG2;
if (cache->cc_key[i] > 0)
keytype = tupdesc->attrs[cache->cc_key[i] - 1]->atttypid;
else
{
if (cache->cc_key[i] != ObjectIdAttributeNumber)
elog(FATAL, "only sys attr supported in caches is OID");
keytype = OIDOID;
}
GetCCHashEqFuncs(keytype,
&cache->cc_hashfunc[i],
&eqfunc);
cache->cc_isname[i] = (keytype == NAMEOID);
/*
* Do equality-function lookup (we assume this won't need a catalog
* lookup for any supported type)
*/
fmgr_info_cxt(eqfunc,
&cache->cc_skey[i].sk_func,
CacheMemoryContext);
/* Initialize sk_attno suitably for HeapKeyTest() and heap scans */
cache->cc_skey[i].sk_attno = cache->cc_key[i];
/* Fill in sk_strategy as well --- always standard equality */
cache->cc_skey[i].sk_strategy = BTEqualStrategyNumber;
cache->cc_skey[i].sk_subtype = InvalidOid;
CACHE4_elog(DEBUG2, "CatalogCacheInitializeCache %s %d %p",
cache->cc_relname,
i,
cache);
}
/*
* mark this cache fully initialized
*/
cache->cc_tupdesc = tupdesc;
}
/*
* Workhorse for AlterLanguageOwner variants
*/
static void
AlterLanguageOwner_internal(HeapTuple tup, Relation rel, Oid newOwnerId)
{
Form_pg_language lanForm;
lanForm = (Form_pg_language) GETSTRUCT(tup);
/*
* If the new owner is the same as the existing owner, consider the
* command to have succeeded. This is for dump restoration purposes.
*/
if (lanForm->lanowner != newOwnerId)
{
Datum repl_val[Natts_pg_language];
bool repl_null[Natts_pg_language];
bool repl_repl[Natts_pg_language];
Acl *newAcl;
Datum aclDatum;
bool isNull;
HeapTuple newtuple;
/* Otherwise, must be owner of the existing object */
if (!pg_language_ownercheck(HeapTupleGetOid(tup), GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_LANGUAGE,
NameStr(lanForm->lanname));
/* Must be able to become new owner */
check_is_member_of_role(GetUserId(), newOwnerId);
memset(repl_null, false, sizeof(repl_null));
memset(repl_repl, false, sizeof(repl_repl));
repl_repl[Anum_pg_language_lanowner - 1] = true;
repl_val[Anum_pg_language_lanowner - 1] = ObjectIdGetDatum(newOwnerId);
/*
* Determine the modified ACL for the new owner. This is only
* necessary when the ACL is non-null.
*/
aclDatum = SysCacheGetAttr(LANGNAME, tup,
Anum_pg_language_lanacl,
&isNull);
if (!isNull)
{
newAcl = aclnewowner(DatumGetAclP(aclDatum),
lanForm->lanowner, newOwnerId);
repl_repl[Anum_pg_language_lanacl - 1] = true;
repl_val[Anum_pg_language_lanacl - 1] = PointerGetDatum(newAcl);
}
newtuple = heap_modify_tuple(tup, RelationGetDescr(rel),
repl_val, repl_null, repl_repl);
simple_heap_update(rel, &newtuple->t_self, newtuple);
CatalogUpdateIndexes(rel, newtuple);
heap_freetuple(newtuple);
/* Update owner dependency reference */
changeDependencyOnOwner(LanguageRelationId, HeapTupleGetOid(tup),
newOwnerId);
}
}
/*
* CreateComments --
*
* Create a comment for the specified object descriptor. Inserts a new
* pg_description tuple, or replaces an existing one with the same key.
*
* If the comment given is null or an empty string, instead delete any
* existing comment for the specified key.
*/
void
CreateComments(Oid oid, Oid classoid, int32 subid, char *comment)
{
Relation description;
ScanKeyData skey[3];
SysScanDesc sd;
HeapTuple oldtuple;
HeapTuple newtuple = NULL;
Datum values[Natts_pg_description];
bool nulls[Natts_pg_description];
bool replaces[Natts_pg_description];
int i;
/* Reduce empty-string to NULL case */
if (comment != NULL && strlen(comment) == 0)
comment = NULL;
/* Prepare to form or update a tuple, if necessary */
if (comment != NULL)
{
for (i = 0; i < Natts_pg_description; i++)
{
nulls[i] = false;
replaces[i] = true;
}
values[Anum_pg_description_objoid - 1] = ObjectIdGetDatum(oid);
values[Anum_pg_description_classoid - 1] = ObjectIdGetDatum(classoid);
values[Anum_pg_description_objsubid - 1] = Int32GetDatum(subid);
values[Anum_pg_description_description - 1] = CStringGetTextDatum(comment);
}
/* Use the index to search for a matching old tuple */
ScanKeyInit(&skey[0],
Anum_pg_description_objoid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(oid));
ScanKeyInit(&skey[1],
Anum_pg_description_classoid,
BTEqualStrategyNumber, F_OIDEQ,
ObjectIdGetDatum(classoid));
ScanKeyInit(&skey[2],
Anum_pg_description_objsubid,
BTEqualStrategyNumber, F_INT4EQ,
Int32GetDatum(subid));
description = heap_open(DescriptionRelationId, RowExclusiveLock);
sd = systable_beginscan(description, DescriptionObjIndexId, true,
SnapshotNow, 3, skey);
while ((oldtuple = systable_getnext(sd)) != NULL)
{
/* Found the old tuple, so delete or update it */
if (comment == NULL)
simple_heap_delete(description, &oldtuple->t_self);
else
{
newtuple = heap_modify_tuple(oldtuple, RelationGetDescr(description), values,
nulls, replaces);
simple_heap_update(description, &oldtuple->t_self, newtuple);
}
break; /* Assume there can be only one match */
}
systable_endscan(sd);
/* If we didn't find an old tuple, insert a new one */
if (newtuple == NULL && comment != NULL)
{
newtuple = heap_form_tuple(RelationGetDescr(description),
values, nulls);
simple_heap_insert(description, newtuple);
}
/* Update indexes, if necessary */
if (newtuple != NULL)
{
CatalogUpdateIndexes(description, newtuple);
heap_freetuple(newtuple);
}
/* Done */
heap_close(description, NoLock);
}
/*
* Map a relation's (tablespace, filenode) to a relation's oid and cache the
* result.
*
* Returns InvalidOid if no relation matching the criteria could be found.
*/
Oid
RelidByRelfilenode(Oid reltablespace, Oid relfilenode)
{
RelfilenodeMapKey key;
RelfilenodeMapEntry *entry;
bool found;
SysScanDesc scandesc;
Relation relation;
HeapTuple ntp;
ScanKeyData skey[2];
Oid relid;
if (RelfilenodeMapHash == NULL)
InitializeRelfilenodeMap();
/* pg_class will show 0 when the value is actually MyDatabaseTableSpace */
if (reltablespace == MyDatabaseTableSpace)
reltablespace = 0;
MemSet(&key, 0, sizeof(key));
key.reltablespace = reltablespace;
key.relfilenode = relfilenode;
/*
* Check cache and return entry if one is found. Even if no target
* relation can be found later on we store the negative match and return a
* InvalidOid from cache. That's not really necessary for performance
* since querying invalid values isn't supposed to be a frequent thing,
* but it's basically free.
*/
entry = hash_search(RelfilenodeMapHash, (void *) &key, HASH_FIND, &found);
if (found)
return entry->relid;
/* ok, no previous cache entry, do it the hard way */
/* initialize empty/negative cache entry before doing the actual lookups */
relid = InvalidOid;
if (reltablespace == GLOBALTABLESPACE_OID)
{
/*
* Ok, shared table, check relmapper.
*/
relid = RelationMapFilenodeToOid(relfilenode, true);
}
else
{
/*
* Not a shared table, could either be a plain relation or a
* non-shared, nailed one, like e.g. pg_class.
*/
/* check for plain relations by looking in pg_class */
relation = heap_open(RelationRelationId, AccessShareLock);
/* copy scankey to local copy, it will be modified during the scan */
memcpy(skey, relfilenode_skey, sizeof(skey));
/* set scan arguments */
skey[0].sk_argument = ObjectIdGetDatum(reltablespace);
skey[1].sk_argument = ObjectIdGetDatum(relfilenode);
scandesc = systable_beginscan(relation,
ClassTblspcRelfilenodeIndexId,
true,
NULL,
2,
skey);
found = false;
while (HeapTupleIsValid(ntp = systable_getnext(scandesc)))
{
if (found)
elog(ERROR,
"unexpected duplicate for tablespace %u, relfilenode %u",
reltablespace, relfilenode);
found = true;
#ifdef USE_ASSERT_CHECKING
{
bool isnull;
Oid check;
check = fastgetattr(ntp, Anum_pg_class_reltablespace,
RelationGetDescr(relation),
&isnull);
Assert(!isnull && check == reltablespace);
check = fastgetattr(ntp, Anum_pg_class_relfilenode,
RelationGetDescr(relation),
&isnull);
Assert(!isnull && check == relfilenode);
}
#endif
relid = HeapTupleGetOid(ntp);
}
systable_endscan(scandesc);
heap_close(relation, AccessShareLock);
/* check for tables that are mapped but not shared */
if (!found)
relid = RelationMapFilenodeToOid(relfilenode, false);
}
/*
* Only enter entry into cache now, our opening of pg_class could have
* caused cache invalidations to be executed which would have deleted a
* new entry if we had entered it above.
*/
entry = hash_search(RelfilenodeMapHash, (void *) &key, HASH_ENTER, &found);
if (found)
elog(ERROR, "corrupted hashtable");
entry->relid = relid;
return relid;
}