static int
combine_files (enum comb_command_type command,
struct lexer *lexer, struct dataset *ds)
{
struct comb_proc proc;
bool saw_by = false;
bool saw_sort = false;
struct casereader *active_file = NULL;
char *first_name = NULL;
char *last_name = NULL;
struct taint *taint = NULL;
size_t n_tables = 0;
size_t allocated_files = 0;
size_t i;
proc.files = NULL;
proc.n_files = 0;
proc.dict = dict_create (get_default_encoding ());
proc.output = NULL;
proc.matcher = NULL;
subcase_init_empty (&proc.by_vars);
proc.first = NULL;
proc.last = NULL;
proc.buffered_case = NULL;
proc.prev_BY = NULL;
dict_set_case_limit (proc.dict, dict_get_case_limit (dataset_dict (ds)));
lex_match (lexer, T_SLASH);
for (;;)
{
struct comb_file *file;
enum comb_file_type type;
if (lex_match_id (lexer, "FILE"))
type = COMB_FILE;
else if (command == COMB_MATCH && lex_match_id (lexer, "TABLE"))
{
type = COMB_TABLE;
n_tables++;
}
else
break;
lex_match (lexer, T_EQUALS);
if (proc.n_files >= allocated_files)
proc.files = x2nrealloc (proc.files, &allocated_files,
sizeof *proc.files);
file = &proc.files[proc.n_files++];
file->type = type;
subcase_init_empty (&file->by_vars);
subcase_init_empty (&file->src);
subcase_init_empty (&file->dst);
file->mv = NULL;
file->handle = NULL;
file->dict = NULL;
file->reader = NULL;
file->data = NULL;
file->is_sorted = true;
file->in_name = NULL;
file->in_var = NULL;
if (lex_match (lexer, T_ASTERISK))
{
if (!dataset_has_source (ds))
{
msg (SE, _("Cannot specify the active dataset since none "
"has been defined."));
goto error;
}
if (proc_make_temporary_transformations_permanent (ds))
msg (SE, _("This command may not be used after TEMPORARY when "
"the active dataset is an input source. "
"Temporary transformations will be made permanent."));
file->dict = dict_clone (dataset_dict (ds));
}
else
{
file->handle = fh_parse (lexer, FH_REF_FILE, dataset_session (ds));
if (file->handle == NULL)
goto error;
file->reader = any_reader_open (file->handle, NULL, &file->dict);
if (file->reader == NULL)
goto error;
}
while (lex_match (lexer, T_SLASH))
if (lex_match_id (lexer, "RENAME"))
{
if (!parse_dict_rename (lexer, file->dict))
goto error;
}
else if (lex_match_id (lexer, "IN"))
{
lex_match (lexer, T_EQUALS);
if (lex_token (lexer) != T_ID)
{
lex_error (lexer, NULL);
goto error;
}
if (file->in_name)
{
msg (SE, _("Multiple IN subcommands for a single FILE or "
"TABLE."));
goto error;
}
file->in_name = xstrdup (lex_tokcstr (lexer));
lex_get (lexer);
}
else if (lex_match_id (lexer, "SORT"))
{
file->is_sorted = false;
saw_sort = true;
}
if (!merge_dictionary (proc.dict, file))
goto error;
}
while (lex_token (lexer) != T_ENDCMD)
{
if (lex_match (lexer, T_BY))
{
const struct variable **by_vars;
size_t i;
bool ok;
if (saw_by)
{
lex_sbc_only_once ("BY");
goto error;
}
saw_by = true;
lex_match (lexer, T_EQUALS);
if (!parse_sort_criteria (lexer, proc.dict, &proc.by_vars,
&by_vars, NULL))
goto error;
ok = true;
for (i = 0; i < proc.n_files; i++)
{
struct comb_file *file = &proc.files[i];
size_t j;
for (j = 0; j < subcase_get_n_fields (&proc.by_vars); j++)
{
const char *name = var_get_name (by_vars[j]);
struct variable *var = dict_lookup_var (file->dict, name);
if (var != NULL)
subcase_add_var (&file->by_vars, var,
subcase_get_direction (&proc.by_vars, j));
else
{
if (file->handle != NULL)
msg (SE, _("File %s lacks BY variable %s."),
fh_get_name (file->handle), name);
else
msg (SE, _("Active dataset lacks BY variable %s."),
name);
ok = false;
}
}
assert (!ok || subcase_conformable (&file->by_vars,
&proc.files[0].by_vars));
}
free (by_vars);
if (!ok)
goto error;
}
else if (command != COMB_UPDATE && lex_match_id (lexer, "FIRST"))
{
if (first_name != NULL)
{
lex_sbc_only_once ("FIRST");
goto error;
}
lex_match (lexer, T_EQUALS);
if (!lex_force_id (lexer))
goto error;
first_name = xstrdup (lex_tokcstr (lexer));
lex_get (lexer);
}
else if (command != COMB_UPDATE && lex_match_id (lexer, "LAST"))
{
if (last_name != NULL)
{
lex_sbc_only_once ("LAST");
goto error;
}
lex_match (lexer, T_EQUALS);
if (!lex_force_id (lexer))
goto error;
last_name = xstrdup (lex_tokcstr (lexer));
lex_get (lexer);
}
else if (lex_match_id (lexer, "MAP"))
{
/* FIXME. */
}
else if (lex_match_id (lexer, "DROP"))
{
if (!parse_dict_drop (lexer, proc.dict))
goto error;
}
else if (lex_match_id (lexer, "KEEP"))
{
if (!parse_dict_keep (lexer, proc.dict))
goto error;
}
else
{
lex_error (lexer, NULL);
goto error;
}
if (!lex_match (lexer, T_SLASH) && lex_token (lexer) != T_ENDCMD)
{
lex_end_of_command (lexer);
goto error;
}
}
if (!saw_by)
{
if (command == COMB_UPDATE)
{
lex_sbc_missing ("BY");
goto error;
}
if (n_tables)
{
msg (SE, _("BY is required when %s is specified."), "TABLE");
goto error;
}
if (saw_sort)
{
msg (SE, _("BY is required when %s is specified."), "SORT");
goto error;
}
}
/* Add IN, FIRST, and LAST variables to master dictionary. */
for (i = 0; i < proc.n_files; i++)
{
struct comb_file *file = &proc.files[i];
if (!create_flag_var ("IN", file->in_name, proc.dict, &file->in_var))
goto error;
}
if (!create_flag_var ("FIRST", first_name, proc.dict, &proc.first)
|| !create_flag_var ("LAST", last_name, proc.dict, &proc.last))
goto error;
dict_delete_scratch_vars (proc.dict);
dict_compact_values (proc.dict);
/* Set up mapping from each file's variables to master
variables. */
for (i = 0; i < proc.n_files; i++)
{
struct comb_file *file = &proc.files[i];
size_t src_var_cnt = dict_get_var_cnt (file->dict);
size_t j;
file->mv = xnmalloc (src_var_cnt, sizeof *file->mv);
for (j = 0; j < src_var_cnt; j++)
{
struct variable *src_var = dict_get_var (file->dict, j);
struct variable *dst_var = dict_lookup_var (proc.dict,
var_get_name (src_var));
if (dst_var != NULL)
{
size_t n = subcase_get_n_fields (&file->src);
file->mv[n] = var_get_missing_values (src_var);
subcase_add_var (&file->src, src_var, SC_ASCEND);
subcase_add_var (&file->dst, dst_var, SC_ASCEND);
}
}
}
proc.output = autopaging_writer_create (dict_get_proto (proc.dict));
taint = taint_clone (casewriter_get_taint (proc.output));
/* Set up case matcher. */
proc.matcher = case_matcher_create ();
for (i = 0; i < proc.n_files; i++)
{
struct comb_file *file = &proc.files[i];
if (file->reader == NULL)
{
if (active_file == NULL)
{
proc_discard_output (ds);
file->reader = active_file = proc_open_filtering (ds, false);
}
else
file->reader = casereader_clone (active_file);
}
if (!file->is_sorted)
file->reader = sort_execute (file->reader, &file->by_vars);
taint_propagate (casereader_get_taint (file->reader), taint);
file->data = casereader_read (file->reader);
if (file->type == COMB_FILE)
case_matcher_add_input (proc.matcher, &file->by_vars,
&file->data, &file->is_minimal);
}
if (command == COMB_ADD)
execute_add_files (&proc);
else if (command == COMB_MATCH)
execute_match_files (&proc);
else if (command == COMB_UPDATE)
execute_update (&proc);
else
NOT_REACHED ();
case_matcher_destroy (proc.matcher);
proc.matcher = NULL;
close_all_comb_files (&proc);
if (active_file != NULL)
proc_commit (ds);
dataset_set_dict (ds, proc.dict);
dataset_set_source (ds, casewriter_make_reader (proc.output));
proc.dict = NULL;
proc.output = NULL;
free_comb_proc (&proc);
free (first_name);
free (last_name);
return taint_destroy (taint) ? CMD_SUCCESS : CMD_CASCADING_FAILURE;
error:
if (active_file != NULL)
proc_commit (ds);
free_comb_proc (&proc);
taint_destroy (taint);
free (first_name);
free (last_name);
return CMD_CASCADING_FAILURE;
}
/* MAIN */
int
main( int argc, char **argv )
{
size_t results[] = {4255, 142803, 46838, 0}, i = 0,
output = 0;
const size_t b = 5;
const char *count_bytes_files[] = {
"/data/orig_breaks.txt", "/data/spawner_output.txt",
"/data/zh_romance_of_three_kingdoms.txt", ""},
*source_dir = getenv("SOURCE");
char buffer[BUFSIZ], tinybuffer[b], *locale = "root", *encoding = "UTF-8";
OChar obuffer[BUFSIZ], otinybuffer[b];
struct stat sts;
time_t now;
FILE *f;
init_oly(argv[0], TEST_PKGDATADIR, encoding, locale);
if (source_dir == NULL)
{
fprintf(stderr, "requires SOURCE environment variable, supplied by runtest. Exiting...\n");
exit(EXIT_FAILURE);
}
plan(16);
diag("----- Testing oly_timestamp function. -----");
time(&now);
is_double(((double)( now*1000)), oly_timestamp(), 1000.0, "Two times should be close enough.");
diag("----- Testing count_file_bytes function. -----");
for (i = 0; (results[i] != 0); i++) {
strcpy(buffer, source_dir);
strcat(buffer, count_bytes_files[i]);
if (stat(buffer, &sts) == -1 && errno == ENOENT)
{
printf ("The file %s doesn't exist...\n", buffer);
}
f = fopen(buffer, "r");
assert( count_file_bytes( f, &output ) == OLY_OKAY ) ;
is_int(results[i], output, "File: %s", count_bytes_files[i]);
fclose(f);
}
diag("----- Testing get_default_locale and get_default_encoding function. -----");
is_unicode_string( u"root",
get_default_locale(), "For this test, default locale should be root.");
is_unicode_string( u"UTF-8",
get_default_encoding(), "For this test, default encoding should be UTF-8.");
diag("----- Testing char_default_locale and char_default_encoding function. -----");
is_string( "root",
char_default_locale(), "For this test, default locale should be root.");
is_string( "UTF-8",
char_default_encoding(), "For this test, default encoding should be UTF-8.");
diag("----- ostr_to_cstr and cstr_to_ostr -----");
is_unicode_string( u"Lorum ipsum etc...",
cstr_to_ostr( obuffer, BUFSIZ, "Lorum ipsum etc..."),
"Char to ochar with space, BUFSIZ = %i", BUFSIZ);
is_unicode_string( u"Loru",
cstr_to_ostr( otinybuffer, b, "Lorum ipsum etc..." ),
"Char to ochar with tiny buffer of %i characters.", (int)b );
is_unicode_string( u"",
cstr_to_ostr( otinybuffer, b, "" ),
"Char to ochar With empty string" );
is_unicode_string( u"",
cstr_to_ostr( otinybuffer, b, NULL ),
"Char to ochar from null" );
is_string( "Back the other way!",
ostr_to_cstr( buffer, BUFSIZ, u"Back the other way!" ),
"OChar to char, BUFSIZ = %i", BUFSIZ );
is_string( "five",
ostr_to_cstr( tinybuffer, b, u"five characters is all that fits." ),
"OChar to char, buffer of %i OChars", (int)b );
is_string( "",
ostr_to_cstr( tinybuffer, b, NULL ),
"Char to ochar a NULL from string." );
is_string( "",
ostr_to_cstr( tinybuffer, b, u"" ),
"Char to ochar with an empty string." );
exit(EXIT_SUCCESS);
}
