PQRYRES CSVColumns(PGLOBAL g, const char *fn, char sep, char q,
int hdr, int mxr, bool info)
{
static int buftyp[] = {TYPE_STRING, TYPE_SHORT, TYPE_STRING,
TYPE_INT, TYPE_INT, TYPE_SHORT};
static XFLD fldtyp[] = {FLD_NAME, FLD_TYPE, FLD_TYPENAME,
FLD_PREC, FLD_LENGTH, FLD_SCALE};
static unsigned int length[] = {6, 6, 8, 10, 10, 6};
char *p, *colname[MAXCOL], dechar, filename[_MAX_PATH], buf[4096];
int i, imax, hmax, n, nerr, phase, blank, digit, dec, type;
int ncol = sizeof(buftyp) / sizeof(int);
int num_read = 0, num_max = 10000000; // Statistics
int len[MAXCOL], typ[MAXCOL], prc[MAXCOL];
FILE *infile;
PQRYRES qrp;
PCOLRES crp;
if (info) {
imax = hmax = 0;
length[0] = 128;
goto skipit;
} // endif info
// num_max = atoi(p+1); // Max num of record to test
#if defined(WIN32)
if (sep == ',' || strnicmp(setlocale(LC_NUMERIC, NULL), "French", 6))
dechar = '.';
else
dechar = ',';
#else // !WIN32
dechar = '.';
#endif // !WIN32
if (trace)
htrc("File %s sep=%c q=%c hdr=%d mxr=%d\n",
SVP(fn), sep, q, hdr, mxr);
if (!fn) {
strcpy(g->Message, MSG(MISSING_FNAME));
return NULL;
} // endif fn
imax = hmax = nerr = 0;
mxr = MY_MAX(0, mxr);
for (i = 0; i < MAXCOL; i++) {
colname[i] = NULL;
len[i] = 0;
typ[i] = TYPE_UNKNOWN;
prc[i] = 0;
} // endfor i
/*********************************************************************/
/* Open the input file. */
/*********************************************************************/
PlugSetPath(filename, fn, PlgGetDataPath(g));
if (!(infile= global_fopen(g, MSGID_CANNOT_OPEN, filename, "r")))
return NULL;
if (hdr) {
/*******************************************************************/
/* Make the column names from the first line. */
/*******************************************************************/
phase = 0;
if (fgets(buf, sizeof(buf), infile)) {
n = strlen(buf) + 1;
buf[n - 2] = '\0';
#if defined(UNIX)
// The file can be imported from Windows
if (buf[n - 3] == '\r')
buf[n - 3] = 0;
#endif // UNIX
p = (char*)PlugSubAlloc(g, NULL, n);
memcpy(p, buf, n);
//skip leading blanks
for (; *p == ' '; p++) ;
if (q && *p == q) {
// Header is quoted
p++;
phase = 1;
} // endif q
colname[0] = p;
} else {
sprintf(g->Message, MSG(FILE_IS_EMPTY), fn);
goto err;
} // endif's
for (i = 1; *p; p++)
if (phase == 1 && *p == q) {
*p = '\0';
phase = 0;
} else if (*p == sep && !phase) {
*p = '\0';
//skip leading blanks
for (; *(p+1) == ' '; p++) ;
if (q && *(p+1) == q) {
// Header is quoted
p++;
phase = 1;
} // endif q
colname[i++] = p + 1;
} // endif sep
num_read++;
imax = hmax = i;
for (i = 0; i < hmax; i++)
length[0] = MY_MAX(length[0], strlen(colname[i]));
} // endif hdr
for (num_read++; num_read <= num_max; num_read++) {
/*******************************************************************/
/* Now start the reading process. Read one line. */
/*******************************************************************/
if (fgets(buf, sizeof(buf), infile)) {
n = strlen(buf);
buf[n - 1] = '\0';
#if defined(UNIX)
// The file can be imported from Windows
if (buf[n - 2] == '\r')
buf[n - 2] = 0;
#endif // UNIX
} else if (feof(infile)) {
sprintf(g->Message, MSG(EOF_AFTER_LINE), num_read -1);
break;
} else {
sprintf(g->Message, MSG(ERR_READING_REC), num_read, fn);
goto err;
} // endif's
/*******************************************************************/
/* Make the test for field lengths. */
/*******************************************************************/
i = n = phase = blank = digit = dec = 0;
for (p = buf; *p; p++)
if (*p == sep) {
if (phase != 1) {
if (i == MAXCOL - 1) {
sprintf(g->Message, MSG(TOO_MANY_FIELDS), num_read, fn);
goto err;
} // endif i
if (n) {
len[i] = MY_MAX(len[i], n);
type = (digit || (dec && n == 1)) ? TYPE_STRING
: (dec) ? TYPE_DOUBLE : TYPE_INT;
typ[i] = MY_MIN(type, typ[i]);
prc[i] = MY_MAX((typ[i] == TYPE_DOUBLE) ? (dec - 1) : 0, prc[i]);
} // endif n
i++;
n = phase = blank = digit = dec = 0;
} else // phase == 1
n++;
} else if (*p == ' ') {
if (phase < 2)
n++;
if (blank)
digit = 1;
} else if (*p == q) {
if (phase == 0) {
if (blank)
if (++nerr > mxr) {
sprintf(g->Message, MSG(MISPLACED_QUOTE), num_read);
goto err;
} else
goto skip;
n = 0;
phase = digit = 1;
} else if (phase == 1) {
if (*(p+1) == q) {
// This is currently not implemented for CSV tables
// if (++nerr > mxr) {
// sprintf(g->Message, MSG(QUOTE_IN_QUOTE), num_read);
// goto err;
// } else
// goto skip;
p++;
n++;
} else
phase = 2;
} else if (++nerr > mxr) { // phase == 2
sprintf(g->Message, MSG(MISPLACED_QUOTE), num_read);
goto err;
} else
goto skip;
} else {
if (phase == 2)
if (++nerr > mxr) {
sprintf(g->Message, MSG(MISPLACED_QUOTE), num_read);
goto err;
} else
goto skip;
// isdigit cannot be used here because of debug assert
if (!strchr("0123456789", *p)) {
if (!digit && *p == dechar)
dec = 1; // Decimal point found
else if (blank || !(*p == '-' || *p == '+'))
digit = 1;
} else if (dec)
dec++; // More decimals
n++;
blank = 1;
} // endif's *p
if (phase == 1)
if (++nerr > mxr) {
sprintf(g->Message, MSG(UNBALANCE_QUOTE), num_read);
goto err;
} else
goto skip;
if (n) {
len[i] = MY_MAX(len[i], n);
type = (digit || n == 0 || (dec && n == 1)) ? TYPE_STRING
: (dec) ? TYPE_DOUBLE : TYPE_INT;
typ[i] = MY_MIN(type, typ[i]);
prc[i] = MY_MAX((typ[i] == TYPE_DOUBLE) ? (dec - 1) : 0, prc[i]);
} // endif n
imax = MY_MAX(imax, i+1);
skip: ; // Skip erroneous line
} // endfor num_read
if (trace) {
htrc("imax=%d Lengths:", imax);
for (i = 0; i < imax; i++)
htrc(" %d", len[i]);
htrc("\n");
} // endif trace
fclose(infile);
skipit:
if (trace)
htrc("CSVColumns: imax=%d hmax=%d len=%d\n",
imax, hmax, length[0]);
/*********************************************************************/
/* Allocate the structures used to refer to the result set. */
/*********************************************************************/
qrp = PlgAllocResult(g, ncol, imax, IDS_COLUMNS + 3,
buftyp, fldtyp, length, false, false);
if (info || !qrp)
return qrp;
qrp->Nblin = imax;
/*********************************************************************/
/* Now get the results into blocks. */
/*********************************************************************/
for (i = 0; i < imax; i++) {
if (i >= hmax) {
sprintf(buf, "COL%.3d", i+1);
p = buf;
} else
p = colname[i];
if (typ[i] == TYPE_UNKNOWN) // Void column
typ[i] = TYPE_STRING;
crp = qrp->Colresp; // Column Name
crp->Kdata->SetValue(p, i);
crp = crp->Next; // Data Type
crp->Kdata->SetValue(typ[i], i);
crp = crp->Next; // Type Name
crp->Kdata->SetValue(GetTypeName(typ[i]), i);
crp = crp->Next; // Precision
crp->Kdata->SetValue(len[i], i);
crp = crp->Next; // Length
crp->Kdata->SetValue(len[i], i);
crp = crp->Next; // Scale (precision)
crp->Kdata->SetValue(prc[i], i);
} // endfor i
/*********************************************************************/
/* Return the result pointer for use by GetData routines. */
/*********************************************************************/
return qrp;
err:
fclose(infile);
return NULL;
} // end of CSVCColumns
PQRYRES DBFColumns(PGLOBAL g, const char *fn, BOOL info)
{
int buftyp[] = {TYPE_STRING, TYPE_SHORT, TYPE_STRING,
TYPE_INT, TYPE_INT, TYPE_SHORT};
XFLD fldtyp[] = {FLD_NAME, FLD_TYPE, FLD_TYPENAME,
FLD_PREC, FLD_LENGTH, FLD_SCALE};
unsigned int length[] = {11, 6, 8, 10, 10, 6};
char buf[2], filename[_MAX_PATH];
int ncol = sizeof(buftyp) / sizeof(int);
int rc, type, len, field, fields;
BOOL bad;
DBFHEADER mainhead;
DESCRIPTOR thisfield;
FILE *infile = NULL;
PQRYRES qrp;
PCOLRES crp;
if (trace)
htrc("DBFColumns: File %s\n", SVP(fn));
if (!info) {
if (!fn) {
strcpy(g->Message, MSG(MISSING_FNAME));
return NULL;
} // endif fn
/************************************************************************/
/* Open the input file. */
/************************************************************************/
PlugSetPath(filename, fn, PlgGetDataPath(g));
if (!(infile= global_fopen(g, MSGID_CANNOT_OPEN, filename, "rb")))
return NULL;
/************************************************************************/
/* Get the first 32 bytes of the header. */
/************************************************************************/
if ((rc = dbfhead(g, infile, filename, &mainhead)) == RC_FX) {
fclose(infile);
return NULL;
} // endif dbfhead
/************************************************************************/
/* Allocate the structures used to refer to the result set. */
/************************************************************************/
fields = mainhead.Fields;
} else
fields = 0;
qrp = PlgAllocResult(g, ncol, fields, IDS_COLUMNS + 3,
buftyp, fldtyp, length, true, false);
if (info || !qrp) {
if (infile)
fclose(infile);
return qrp;
} // endif info
if (trace) {
htrc("Structure of %s\n", filename);
htrc("headlen=%hd reclen=%hd degree=%d\n",
mainhead.Headlen, mainhead.Reclen, fields);
htrc("flags(iem)=%d,%d,%d cp=%d\n", mainhead.Incompleteflag,
mainhead.Encryptflag, mainhead.Mdxflag, mainhead.Language);
htrc("%hd records, last changed %02d/%02d/%d\n",
mainhead.Records, mainhead.Filedate[1], mainhead.Filedate[2],
mainhead.Filedate[0] + (mainhead.Filedate[0] <= 30) ? 2000 : 1900);
htrc("Field Type Offset Len Dec Set Mdx\n");
} // endif trace
buf[1] = '\0';
/**************************************************************************/
/* Do it field by field. We are at byte 32 of file. */
/**************************************************************************/
for (field = 0; field < fields; field++) {
bad = FALSE;
if (fread(&thisfield, HEADLEN, 1, infile) != 1) {
sprintf(g->Message, MSG(ERR_READING_REC), field+1, fn);
goto err;
} else
len = thisfield.Length;
if (trace)
htrc("%-11s %c %6ld %3d %2d %3d %3d\n",
thisfield.Name, thisfield.Type, thisfield.Offset, len,
thisfield.Decimals, thisfield.Setfield, thisfield.Mdxfield);
/************************************************************************/
/* Now get the results into blocks. */
/************************************************************************/
switch (thisfield.Type) {
case 'C': // Characters
case 'L': // Logical 'T' or 'F'
type = TYPE_STRING;
break;
case 'N':
type = (thisfield.Decimals) ? TYPE_DOUBLE
: (len > 10) ? TYPE_BIGINT : TYPE_INT;
break;
case 'F':
type = TYPE_DOUBLE;
break;
case 'D':
type = TYPE_DATE; // Is this correct ???
break;
default:
if (!info) {
sprintf(g->Message, MSG(BAD_DBF_TYPE), thisfield.Type);
goto err;
} // endif info
type = TYPE_ERROR;
bad = TRUE;
} // endswitch Type
crp = qrp->Colresp; // Column Name
crp->Kdata->SetValue(thisfield.Name, field);
crp = crp->Next; // Data Type
crp->Kdata->SetValue((int)type, field);
crp = crp->Next; // Type Name
if (bad) {
buf[0] = thisfield.Type;
crp->Kdata->SetValue(buf, field);
} else
crp->Kdata->SetValue(GetTypeName(type), field);
crp = crp->Next; // Precision
crp->Kdata->SetValue((int)thisfield.Length, field);
crp = crp->Next; // Length
crp->Kdata->SetValue((int)thisfield.Length, field);
crp = crp->Next; // Scale (precision)
crp->Kdata->SetValue((int)thisfield.Decimals, field);
} // endfor field
qrp->Nblin = field;
fclose(infile);
#if 0
if (info) {
/************************************************************************/
/* Prepare return message for dbfinfo command. */
/************************************************************************/
char buf[64];
sprintf(buf,
"Ver=%02x ncol=%hu nlin=%u lrecl=%hu headlen=%hu date=%02d/%02d/%02d",
mainhead.Version, fields, mainhead.Records, mainhead.Reclen,
mainhead.Headlen, mainhead.Filedate[0], mainhead.Filedate[1],
mainhead.Filedate[2]);
strcat(g->Message, buf);
} // endif info
#endif // 0
/**************************************************************************/
/* Return the result pointer for use by GetData routines. */
/**************************************************************************/
return qrp;
err:
fclose(infile);
return NULL;
} // end of DBFColumns