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 lex_match_id ("TYPE");
113 if (lex_match_id ("MIXED"))
114 fty->type = FTY_MIXED;
115 else if (lex_match_id ("GROUPED"))
117 fty->type = FTY_GROUPED;
123 else if (lex_match_id ("NESTED"))
124 fty->type = FTY_NESTED;
127 msg (SE, _("MIXED, GROUPED, or NESTED expected."));
133 if (lex_match_id ("FILE"))
136 fty->handle = fh_parse_file_handle ();
140 else if (lex_match_id ("RECORD"))
143 if (!parse_col_spec (&fty->record, "####RECD"))
146 else if (lex_match_id ("CASE"))
148 if (fty->type == FTY_MIXED)
150 msg (SE, _("The CASE subcommand is not valid on FILE TYPE "
156 if (!parse_col_spec (&fty->case_sbc, "####CASE"))
159 else if (lex_match_id ("WILD"))
162 if (lex_match_id ("WARN"))
164 else if (lex_match_id ("NOWARN"))
168 msg (SE, _("WARN or NOWARN expected after WILD."));
172 else if (lex_match_id ("DUPLICATE"))
174 if (fty->type == FTY_MIXED)
176 msg (SE, _("The DUPLICATE subcommand is not valid on "
177 "FILE TYPE MIXED."));
182 if (lex_match_id ("WARN"))
184 else if (lex_match_id ("NOWARN"))
186 else if (lex_match_id ("CASE"))
188 if (fty->type != FTY_NESTED)
190 msg (SE, _("DUPLICATE=CASE is only valid on "
191 "FILE TYPE NESTED."));
199 msg (SE, _("WARN%s expected after DUPLICATE."),
200 (fty->type == FTY_NESTED ? _(", NOWARN, or CASE")
205 else if (lex_match_id ("MISSING"))
207 if (fty->type == FTY_MIXED)
209 msg (SE, _("The MISSING subcommand is not valid on "
210 "FILE TYPE MIXED."));
215 if (lex_match_id ("NOWARN"))
217 else if (lex_match_id ("WARN"))
221 msg (SE, _("WARN or NOWARN after MISSING."));
225 else if (lex_match_id ("ORDERED"))
227 if (fty->type != FTY_GROUPED)
229 msg (SE, _("ORDERED is only valid on FILE TYPE GROUPED."));
234 if (lex_match_id ("YES"))
236 else if (lex_match_id ("NO"))
240 msg (SE, _("YES or NO expected after ORDERED."));
246 lex_error (_("while expecting a valid subcommand"));
251 if (fty->record.name[0] == 0)
253 msg (SE, _("The required RECORD subcommand was not present."));
257 if (fty->type == FTY_GROUPED)
259 if (fty->case_sbc.name[0] == 0)
261 msg (SE, _("The required CASE subcommand was not present."));
265 if (!strcmp (fty->case_sbc.name, fty->record.name))
267 msg (SE, _("CASE and RECORD must specify different variable "
273 default_handle = fty->handle;
275 create_col_var (&fty->record);
276 if (fty->case_sbc.name[0])
277 create_col_var (&fty->case_sbc);
278 vfm_source = create_case_source (&file_type_source_class, default_dict, fty);
287 /* Creates a variable with attributes specified by struct col_spec C, and
288 stores it into C->V. */
290 create_col_var (struct col_spec *c)
294 if (formats[c->fmt].cat & FCAT_STRING)
298 c->v = dict_create_var (default_dict, c->name, width);
301 /* Parses variable, column, type specifications for a variable. */
303 parse_col_spec (struct col_spec *c, const char *def_name)
305 struct fmt_spec spec;
310 strcpy (c->name, tokid);
314 strcpy (c->name, def_name);
317 if (!lex_force_int ())
319 c->fc = lex_integer ();
322 msg (SE, _("Column value must be positive."));
328 lex_negative_to_dash ();
331 if (!lex_force_int ())
333 c->nc = lex_integer ();
338 msg (SE, _("Ending column precedes beginning column."));
347 /* Format specifier. */
351 if (!lex_force_id ())
353 c->fmt = parse_format_specifier_name (&cp, 0);
358 msg (SE, _("Bad format specifier name."));
362 if (!lex_force_match (')'))
371 return check_input_specifier (&spec);
376 /* Parse the RECORD TYPE command. */
378 cmd_record_type (void)
380 struct file_type_pgm *fty;
381 struct record_type *rct;
383 /* Make sure we're inside a FILE TYPE structure. */
384 if (pgm_state != STATE_INPUT
385 || !case_source_is_class (vfm_source, &file_type_source_class))
387 msg (SE, _("This command may only appear within a "
388 "FILE TYPE/END FILE TYPE structure."));
392 fty = vfm_source->aux;
394 /* Initialize the record_type structure. */
395 rct = xmalloc (sizeof *rct);
399 rct->flags |= RCT_DUPLICATE;
401 rct->flags |= RCT_MISSING;
405 if (fty->case_sbc.name[0])
406 rct->case_sbc = fty->case_sbc;
408 if (fty->recs_tail && (fty->recs_tail->flags & RCT_OTHER))
410 msg (SE, _("OTHER may appear only on the last RECORD TYPE command."));
416 fty->recs_tail->lt = n_trns - 1;
417 if (!(fty->recs_tail->flags & RCT_SKIP)
418 && fty->recs_tail->ft == fty->recs_tail->lt)
420 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
421 "for above RECORD TYPE."));
426 lex_match_id ("RECORD");
427 lex_match_id ("TYPE");
429 /* Parse record type values. */
430 if (lex_match_id ("OTHER"))
431 rct->flags |= RCT_OTHER;
436 while (token == T_NUM || token == T_STRING)
441 rct->v = xrealloc (rct->v, mv * sizeof *rct->v);
444 if (formats[fty->record.fmt].cat & FCAT_STRING)
446 if (!lex_force_string ())
448 rct->v[rct->nv].c = xmalloc (fty->record.nc + 1);
449 st_bare_pad_copy (rct->v[rct->nv].c, ds_value (&tokstr),
454 if (!lex_force_num ())
456 rct->v[rct->nv].f = tokval;
465 /* Parse the rest of the subcommands. */
468 if (lex_match_id ("SKIP"))
469 rct->flags |= RCT_SKIP;
470 else if (lex_match_id ("CASE"))
472 if (fty->type == FTY_MIXED)
474 msg (SE, _("The CASE subcommand is not allowed on "
475 "the RECORD TYPE command for FILE TYPE MIXED."));
480 if (!parse_col_spec (&rct->case_sbc, ""))
482 if (rct->case_sbc.name[0])
484 msg (SE, _("No variable name may be specified for the "
485 "CASE subcommand on RECORD TYPE."));
489 if ((formats[rct->case_sbc.fmt].cat ^ formats[fty->case_sbc.fmt].cat)
492 msg (SE, _("The CASE column specification on RECORD TYPE "
493 "must give a format specifier that is the "
494 "same type as that of the CASE column "
495 "specification given on FILE TYPE."));
499 else if (lex_match_id ("DUPLICATE"))
502 if (lex_match_id ("WARN"))
503 rct->flags |= RCT_DUPLICATE;
504 else if (lex_match_id ("NOWARN"))
505 rct->flags &= ~RCT_DUPLICATE;
508 msg (SE, _("WARN or NOWARN expected on DUPLICATE "
513 else if (lex_match_id ("MISSING"))
516 if (lex_match_id ("WARN"))
517 rct->flags |= RCT_MISSING;
518 else if (lex_match_id ("NOWARN"))
519 rct->flags &= ~RCT_MISSING;
522 msg (SE, _("WARN or NOWARN expected on MISSING subcommand."));
526 else if (lex_match_id ("SPREAD"))
529 if (lex_match_id ("YES"))
530 rct->flags |= RCT_SPREAD;
531 else if (lex_match_id ("NO"))
532 rct->flags &= ~RCT_SPREAD;
535 msg (SE, _("YES or NO expected on SPREAD subcommand."));
541 lex_error (_("while expecting a valid subcommand"));
547 fty->recs_tail = fty->recs_tail->next = xmalloc (sizeof *fty->recs_tail);
549 fty->recs_head = fty->recs_tail = xmalloc (sizeof *fty->recs_tail);
550 memcpy (fty->recs_tail, &rct, sizeof *fty->recs_tail);
555 if (formats[fty->record.fmt].cat & FCAT_STRING)
559 for (i = 0; i < rct->nv; i++)
571 cmd_end_file_type (void)
573 struct file_type_pgm *fty;
575 if (pgm_state != STATE_INPUT
576 || case_source_is_class (vfm_source, &file_type_source_class))
578 msg (SE, _("This command may only appear within a "
579 "FILE TYPE/END FILE TYPE structure."));
582 fty = vfm_source->aux;
583 fty->case_size = dict_get_case_size (default_dict);
585 lex_match_id ("TYPE");
589 fty->recs_tail->lt = n_trns - 1;
590 if (!(fty->recs_tail->flags & RCT_SKIP)
591 && fty->recs_tail->ft == fty->recs_tail->lt)
593 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
594 "on above RECORD TYPE."));
600 msg (SE, _("No commands between FILE TYPE and END FILE TYPE."));
606 return lex_end_of_command ();
609 /* Come here on discovering catastrophic error. */
611 discard_variables ();
615 /* FILE TYPE runtime. */
617 /*static void read_from_file_type_mixed(void);
618 static void read_from_file_type_grouped(void);
619 static void read_from_file_type_nested(void); */
621 /* Reads any number of cases into case C and calls write_case()
622 for each one. Compare data-list.c:read_from_data_list. */
624 file_type_source_read (struct case_source *source,
626 write_case_func *write_case UNUSED,
627 write_case_data wc_data UNUSED)
629 struct file_type_pgm *fty = source->aux;
633 struct fmt_spec format;
635 dfm_push (fty->handle);
637 format.type = fty->record.fmt;
638 format.w = fty->record.nc;
640 while (NULL != (line = dfm_get_record (fty->handle, &len)))
642 struct record_type *iter;
646 if (formats[fty->record.fmt].cat & FCAT_STRING)
650 v.c = c->data[fty->record.v->fv].s;
652 data_in_finite_line (&di, line, len,
653 fty->record.fc, fty->record.fc + fty->record.nc);
654 di.v = (union value *) v.c;
656 di.f1 = fty->record.fc;
660 for (iter = fty->recs_head; iter; iter = iter->next)
662 if (iter->flags & RCT_OTHER)
664 for (i = 0; i < iter->nv; i++)
665 if (!memcmp (iter->v[i].c, v.c, fty->record.nc))
669 msg (SW, _("Unknown record type \"%.*s\"."), fty->record.nc, v.c);
675 data_in_finite_line (&di, line, len,
676 fty->record.fc, fty->record.fc + fty->record.nc);
679 di.f1 = fty->record.fc;
683 memcpy (&c->data[fty->record.v->fv].f, &v.f, sizeof v.f);
684 for (iter = fty->recs_head; iter; iter = iter->next)
686 if (iter->flags & RCT_OTHER)
688 for (i = 0; i < iter->nv; i++)
689 if (iter->v[i].f == v.f)
693 msg (SW, _("Unknown record type %g."), v.f);
695 dfm_fwd_record (fty->handle);
699 /* Arrive here if there is a matching record_type, which is in
701 dfm_fwd_record (fty->handle);
706 case FTY_MIXED: read_from_file_type_mixed(); break;
707 case FTY_GROUPED: read_from_file_type_grouped(); break;
708 case FTY_NESTED: read_from_file_type_nested(); break;
712 dfm_pop (fty->handle);
716 file_type_source_destroy (struct case_source *source)
718 struct file_type_pgm *fty = source->aux;
719 struct record_type *iter, *next;
721 cancel_transformations ();
722 for (iter = fty->recs_head; iter; iter = next)
729 const struct case_source_class file_type_source_class =
733 file_type_source_read,
734 file_type_source_destroy,