1 /* PSPP - computes sample statistics.
2 Copyright (C) 1997-9, 2000 Free Software Foundation, Inc.
3 Written by Ben Pfaff <blp@gnu.org>.
5 This program is free software; you can redistribute it and/or
6 modify it under the terms of the GNU General Public License as
7 published by the Free Software Foundation; either version 2 of the
8 License, or (at your option) any later version.
10 This program is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program; if not, write to the Free Software
17 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
27 #include "file-handle.h"
34 /* Defines the three types of complex files read by FILE TYPE. */
42 /* Limited variable column specifications. */
45 char name[9]; /* Variable name. */
46 int fc, nc; /* First column (1-based), # of columns. */
47 int fmt; /* Format type. */
48 struct variable *v; /* Variable. */
51 /* RCT_* record type constants. */
54 RCT_OTHER = 001, /* 1=OTHER. */
55 RCT_SKIP = 002, /* 1=SKIP. */
56 RCT_DUPLICATE = 004, /* DUPLICATE: 0=NOWARN, 1=WARN. */
57 RCT_MISSING = 010, /* MISSING: 0=NOWARN, 1=WARN. */
58 RCT_SPREAD = 020 /* SPREAD: 0=NO, 1=YES. */
61 /* Represents a RECORD TYPE command. */
64 struct record_type *next;
65 unsigned flags; /* RCT_* constants. */
66 union value *v; /* Vector of values for this record type. */
67 int nv; /* Length of vector V. */
68 struct col_spec case_sbc; /* CASE subcommand. */
69 int ft, lt; /* First, last transformation index. */
72 /* Represents a FILE TYPE input program. Does not contain a
73 trns_header because it's never submitted as a transformation. */
76 int type; /* One of the FTY_* constants. */
77 struct file_handle *handle; /* File handle of input file. */
78 struct col_spec record; /* RECORD subcommand. */
79 struct col_spec case_sbc; /* CASE subcommand. */
80 int wild; /* 0=NOWARN, 1=WARN. */
81 int duplicate; /* 0=NOWARN, 1=WARN. */
82 int missing; /* 0=NOWARN, 1=WARN, 2=CASE. */
83 int ordered; /* 0=NO, 1=YES. */
84 int had_rec_type; /* 1=Had a RECORD TYPE command.
85 RECORD TYPE must precede the first
87 struct record_type *recs_head; /* List of record types. */
88 struct record_type *recs_tail; /* Last in list of record types. */
89 size_t case_size; /* Case size in bytes. */
92 static int parse_col_spec (struct col_spec *, const char *);
93 static void create_col_var (struct col_spec *c);
95 /* Parses FILE TYPE command. */
99 static struct file_type_pgm *fty;
102 discard_variables ();
104 fty = xmalloc (sizeof *fty);
105 fty->handle = inline_file;
106 fty->record.name[0] = 0;
107 fty->case_sbc.name[0] = 0;
108 fty->wild = fty->duplicate = fty->missing = fty->ordered = 0;
109 fty->had_rec_type = 0;
110 fty->recs_head = fty->recs_tail = NULL;
112 if (lex_match_id ("MIXED"))
113 fty->type = FTY_MIXED;
114 else if (lex_match_id ("GROUPED"))
116 fty->type = FTY_GROUPED;
122 else if (lex_match_id ("NESTED"))
123 fty->type = FTY_NESTED;
126 msg (SE, _("MIXED, GROUPED, or NESTED expected."));
132 if (lex_match_id ("FILE"))
135 fty->handle = fh_parse_file_handle ();
139 else if (lex_match_id ("RECORD"))
142 if (!parse_col_spec (&fty->record, "####RECD"))
145 else if (lex_match_id ("CASE"))
147 if (fty->type == FTY_MIXED)
149 msg (SE, _("The CASE subcommand is not valid on FILE TYPE "
155 if (!parse_col_spec (&fty->case_sbc, "####CASE"))
158 else if (lex_match_id ("WILD"))
161 if (lex_match_id ("WARN"))
163 else if (lex_match_id ("NOWARN"))
167 msg (SE, _("WARN or NOWARN expected after WILD."));
171 else if (lex_match_id ("DUPLICATE"))
173 if (fty->type == FTY_MIXED)
175 msg (SE, _("The DUPLICATE subcommand is not valid on "
176 "FILE TYPE MIXED."));
181 if (lex_match_id ("WARN"))
183 else if (lex_match_id ("NOWARN"))
185 else if (lex_match_id ("CASE"))
187 if (fty->type != FTY_NESTED)
189 msg (SE, _("DUPLICATE=CASE is only valid on "
190 "FILE TYPE NESTED."));
198 msg (SE, _("WARN%s expected after DUPLICATE."),
199 (fty->type == FTY_NESTED ? _(", NOWARN, or CASE")
204 else if (lex_match_id ("MISSING"))
206 if (fty->type == FTY_MIXED)
208 msg (SE, _("The MISSING subcommand is not valid on "
209 "FILE TYPE MIXED."));
214 if (lex_match_id ("NOWARN"))
216 else if (lex_match_id ("WARN"))
220 msg (SE, _("WARN or NOWARN after MISSING."));
224 else if (lex_match_id ("ORDERED"))
226 if (fty->type != FTY_GROUPED)
228 msg (SE, _("ORDERED is only valid on FILE TYPE GROUPED."));
233 if (lex_match_id ("YES"))
235 else if (lex_match_id ("NO"))
239 msg (SE, _("YES or NO expected after ORDERED."));
245 lex_error (_("while expecting a valid subcommand"));
250 if (fty->record.name[0] == 0)
252 msg (SE, _("The required RECORD subcommand was not present."));
256 if (fty->type == FTY_GROUPED)
258 if (fty->case_sbc.name[0] == 0)
260 msg (SE, _("The required CASE subcommand was not present."));
264 if (!strcmp (fty->case_sbc.name, fty->record.name))
266 msg (SE, _("CASE and RECORD must specify different variable "
272 if (!dfm_open_for_reading (fty->handle))
274 default_handle = fty->handle;
276 create_col_var (&fty->record);
277 if (fty->case_sbc.name[0])
278 create_col_var (&fty->case_sbc);
279 vfm_source = create_case_source (&file_type_source_class, default_dict, fty);
288 /* Creates a variable with attributes specified by struct col_spec C, and
289 stores it into C->V. */
291 create_col_var (struct col_spec *c)
295 if (formats[c->fmt].cat & FCAT_STRING)
299 c->v = dict_create_var (default_dict, c->name, width);
302 /* Parses variable, column, type specifications for a variable. */
304 parse_col_spec (struct col_spec *c, const char *def_name)
306 struct fmt_spec spec;
311 strcpy (c->name, tokid);
315 strcpy (c->name, def_name);
318 if (!lex_force_int ())
320 c->fc = lex_integer ();
323 msg (SE, _("Column value must be positive."));
329 lex_negative_to_dash ();
332 if (!lex_force_int ())
334 c->nc = lex_integer ();
339 msg (SE, _("Ending column precedes beginning column."));
348 /* Format specifier. */
352 if (!lex_force_id ())
354 c->fmt = parse_format_specifier_name (&cp, 0);
359 msg (SE, _("Bad format specifier name."));
363 if (!lex_force_match (')'))
372 return check_input_specifier (&spec);
377 /* Parse the RECORD TYPE command. */
379 cmd_record_type (void)
381 struct file_type_pgm *fty;
382 struct record_type *rct;
384 /* Make sure we're inside a FILE TYPE structure. */
385 if (pgm_state != STATE_INPUT
386 || !case_source_is_class (vfm_source, &file_type_source_class))
388 msg (SE, _("This command may only appear within a "
389 "FILE TYPE/END FILE TYPE structure."));
393 fty = vfm_source->aux;
395 /* Initialize the record_type structure. */
396 rct = xmalloc (sizeof *rct);
400 rct->flags |= RCT_DUPLICATE;
402 rct->flags |= RCT_MISSING;
406 if (fty->case_sbc.name[0])
407 rct->case_sbc = fty->case_sbc;
409 if (fty->recs_tail && (fty->recs_tail->flags & RCT_OTHER))
411 msg (SE, _("OTHER may appear only on the last RECORD TYPE command."));
417 fty->recs_tail->lt = n_trns - 1;
418 if (!(fty->recs_tail->flags & RCT_SKIP)
419 && fty->recs_tail->ft == fty->recs_tail->lt)
421 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
422 "for above RECORD TYPE."));
427 /* Parse record type values. */
428 if (lex_match_id ("OTHER"))
429 rct->flags |= RCT_OTHER;
434 while (token == T_NUM || token == T_STRING)
439 rct->v = xrealloc (rct->v, mv * sizeof *rct->v);
442 if (formats[fty->record.fmt].cat & FCAT_STRING)
444 if (!lex_force_string ())
446 rct->v[rct->nv].c = xmalloc (fty->record.nc + 1);
447 st_bare_pad_copy (rct->v[rct->nv].c, ds_value (&tokstr),
452 if (!lex_force_num ())
454 rct->v[rct->nv].f = tokval;
463 /* Parse the rest of the subcommands. */
466 if (lex_match_id ("SKIP"))
467 rct->flags |= RCT_SKIP;
468 else if (lex_match_id ("CASE"))
470 if (fty->type == FTY_MIXED)
472 msg (SE, _("The CASE subcommand is not allowed on "
473 "the RECORD TYPE command for FILE TYPE MIXED."));
478 if (!parse_col_spec (&rct->case_sbc, ""))
480 if (rct->case_sbc.name[0])
482 msg (SE, _("No variable name may be specified for the "
483 "CASE subcommand on RECORD TYPE."));
487 if ((formats[rct->case_sbc.fmt].cat ^ formats[fty->case_sbc.fmt].cat)
490 msg (SE, _("The CASE column specification on RECORD TYPE "
491 "must give a format specifier that is the "
492 "same type as that of the CASE column "
493 "specification given on FILE TYPE."));
497 else if (lex_match_id ("DUPLICATE"))
500 if (lex_match_id ("WARN"))
501 rct->flags |= RCT_DUPLICATE;
502 else if (lex_match_id ("NOWARN"))
503 rct->flags &= ~RCT_DUPLICATE;
506 msg (SE, _("WARN or NOWARN expected on DUPLICATE "
511 else if (lex_match_id ("MISSING"))
514 if (lex_match_id ("WARN"))
515 rct->flags |= RCT_MISSING;
516 else if (lex_match_id ("NOWARN"))
517 rct->flags &= ~RCT_MISSING;
520 msg (SE, _("WARN or NOWARN expected on MISSING subcommand."));
524 else if (lex_match_id ("SPREAD"))
527 if (lex_match_id ("YES"))
528 rct->flags |= RCT_SPREAD;
529 else if (lex_match_id ("NO"))
530 rct->flags &= ~RCT_SPREAD;
533 msg (SE, _("YES or NO expected on SPREAD subcommand."));
539 lex_error (_("while expecting a valid subcommand"));
545 fty->recs_tail = fty->recs_tail->next = xmalloc (sizeof *fty->recs_tail);
547 fty->recs_head = fty->recs_tail = xmalloc (sizeof *fty->recs_tail);
548 memcpy (fty->recs_tail, &rct, sizeof *fty->recs_tail);
553 if (formats[fty->record.fmt].cat & FCAT_STRING)
557 for (i = 0; i < rct->nv; i++)
569 cmd_end_file_type (void)
571 struct file_type_pgm *fty;
573 if (pgm_state != STATE_INPUT
574 || case_source_is_class (vfm_source, &file_type_source_class))
576 msg (SE, _("This command may only appear within a "
577 "FILE TYPE/END FILE TYPE structure."));
580 fty = vfm_source->aux;
581 fty->case_size = dict_get_case_size (default_dict);
585 fty->recs_tail->lt = n_trns - 1;
586 if (!(fty->recs_tail->flags & RCT_SKIP)
587 && fty->recs_tail->ft == fty->recs_tail->lt)
589 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
590 "on above RECORD TYPE."));
596 msg (SE, _("No commands between FILE TYPE and END FILE TYPE."));
602 return lex_end_of_command ();
605 /* Come here on discovering catastrophic error. */
607 discard_variables ();
611 /* FILE TYPE runtime. */
613 /*static void read_from_file_type_mixed(void);
614 static void read_from_file_type_grouped(void);
615 static void read_from_file_type_nested(void); */
617 /* Reads any number of cases into case C and calls write_case()
618 for each one. Compare data-list.c:read_from_data_list. */
620 file_type_source_read (struct case_source *source,
622 write_case_func *write_case UNUSED,
623 write_case_data wc_data UNUSED)
625 struct file_type_pgm *fty = source->aux;
629 struct fmt_spec format;
631 dfm_push (fty->handle);
633 format.type = fty->record.fmt;
634 format.w = fty->record.nc;
636 while (NULL != (line = dfm_get_record (fty->handle, &len)))
638 struct record_type *iter;
642 if (formats[fty->record.fmt].cat & FCAT_STRING)
646 v.c = c->data[fty->record.v->fv].s;
648 data_in_finite_line (&di, line, len,
649 fty->record.fc, fty->record.fc + fty->record.nc);
650 di.v = (union value *) v.c;
652 di.f1 = fty->record.fc;
656 for (iter = fty->recs_head; iter; iter = iter->next)
658 if (iter->flags & RCT_OTHER)
660 for (i = 0; i < iter->nv; i++)
661 if (!memcmp (iter->v[i].c, v.c, fty->record.nc))
665 msg (SW, _("Unknown record type \"%.*s\"."), fty->record.nc, v.c);
671 data_in_finite_line (&di, line, len,
672 fty->record.fc, fty->record.fc + fty->record.nc);
675 di.f1 = fty->record.fc;
679 memcpy (&c->data[fty->record.v->fv].f, &v.f, sizeof v.f);
680 for (iter = fty->recs_head; iter; iter = iter->next)
682 if (iter->flags & RCT_OTHER)
684 for (i = 0; i < iter->nv; i++)
685 if (iter->v[i].f == v.f)
689 msg (SW, _("Unknown record type %g."), v.f);
691 dfm_fwd_record (fty->handle);
695 /* Arrive here if there is a matching record_type, which is in
697 dfm_fwd_record (fty->handle);
702 case FTY_MIXED: read_from_file_type_mixed(); break;
703 case FTY_GROUPED: read_from_file_type_grouped(); break;
704 case FTY_NESTED: read_from_file_type_nested(); break;
708 dfm_pop (fty->handle);
712 file_type_source_destroy (struct case_source *source)
714 struct file_type_pgm *fty = source->aux;
715 struct record_type *iter, *next;
717 cancel_transformations ();
718 for (iter = fty->recs_head; iter; iter = next)
725 const struct case_source_class file_type_source_class =
729 file_type_source_read,
730 file_type_source_destroy,