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. */
91 static int parse_col_spec (struct col_spec *, const char *);
92 static void create_col_var (struct col_spec *c);
94 /* Parses FILE TYPE command. */
98 static struct file_type_pgm *fty;
101 discard_variables ();
103 fty = xmalloc (sizeof *fty);
104 fty->handle = inline_file;
105 fty->record.name[0] = 0;
106 fty->case_sbc.name[0] = 0;
107 fty->wild = fty->duplicate = fty->missing = fty->ordered = 0;
108 fty->had_rec_type = 0;
109 fty->recs_head = fty->recs_tail = NULL;
111 lex_match_id ("TYPE");
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 default_handle = fty->handle;
274 vfm_source = create_case_source (&file_type_source_class, fty);
275 create_col_var (&fty->record);
276 if (fty->case_sbc.name[0])
277 create_col_var (&fty->case_sbc);
286 /* Creates a variable with attributes specified by struct col_spec C, and
287 stores it into C->V. */
289 create_col_var (struct col_spec *c)
293 if (formats[c->fmt].cat & FCAT_STRING)
297 c->v = dict_create_var (default_dict, c->name, width);
300 /* Parses variable, column, type specifications for a variable. */
302 parse_col_spec (struct col_spec *c, const char *def_name)
304 struct fmt_spec spec;
309 strcpy (c->name, tokid);
313 strcpy (c->name, def_name);
316 if (!lex_force_int ())
318 c->fc = lex_integer ();
321 msg (SE, _("Column value must be positive."));
327 lex_negative_to_dash ();
330 if (!lex_force_int ())
332 c->nc = lex_integer ();
337 msg (SE, _("Ending column precedes beginning column."));
346 /* Format specifier. */
350 if (!lex_force_id ())
352 c->fmt = parse_format_specifier_name (&cp, 0);
357 msg (SE, _("Bad format specifier name."));
361 if (!lex_force_match (')'))
370 return check_input_specifier (&spec);
375 /* Parse the RECORD TYPE command. */
377 cmd_record_type (void)
379 struct file_type_pgm *fty;
380 struct record_type *rct;
382 /* Make sure we're inside a FILE TYPE structure. */
383 if (pgm_state != STATE_INPUT
384 || !case_source_is_class (vfm_source, &file_type_source_class))
386 msg (SE, _("This command may only appear within a "
387 "FILE TYPE/END FILE TYPE structure."));
391 fty = vfm_source->aux;
393 /* Initialize the record_type structure. */
394 rct = xmalloc (sizeof *rct);
398 rct->flags |= RCT_DUPLICATE;
400 rct->flags |= RCT_MISSING;
404 if (fty->case_sbc.name[0])
405 rct->case_sbc = fty->case_sbc;
407 if (fty->recs_tail && (fty->recs_tail->flags & RCT_OTHER))
409 msg (SE, _("OTHER may appear only on the last RECORD TYPE command."));
415 fty->recs_tail->lt = n_trns - 1;
416 if (!(fty->recs_tail->flags & RCT_SKIP)
417 && fty->recs_tail->ft == fty->recs_tail->lt)
419 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
420 "for above RECORD TYPE."));
425 lex_match_id ("RECORD");
426 lex_match_id ("TYPE");
428 /* Parse record type values. */
429 if (lex_match_id ("OTHER"))
430 rct->flags |= RCT_OTHER;
435 while (token == T_NUM || token == T_STRING)
440 rct->v = xrealloc (rct->v, mv * sizeof *rct->v);
443 if (formats[fty->record.fmt].cat & FCAT_STRING)
445 if (!lex_force_string ())
447 rct->v[rct->nv].c = xmalloc (fty->record.nc + 1);
448 st_bare_pad_copy (rct->v[rct->nv].c, ds_value (&tokstr),
453 if (!lex_force_num ())
455 rct->v[rct->nv].f = tokval;
464 /* Parse the rest of the subcommands. */
467 if (lex_match_id ("SKIP"))
468 rct->flags |= RCT_SKIP;
469 else if (lex_match_id ("CASE"))
471 if (fty->type == FTY_MIXED)
473 msg (SE, _("The CASE subcommand is not allowed on "
474 "the RECORD TYPE command for FILE TYPE MIXED."));
479 if (!parse_col_spec (&rct->case_sbc, ""))
481 if (rct->case_sbc.name[0])
483 msg (SE, _("No variable name may be specified for the "
484 "CASE subcommand on RECORD TYPE."));
488 if ((formats[rct->case_sbc.fmt].cat ^ formats[fty->case_sbc.fmt].cat)
491 msg (SE, _("The CASE column specification on RECORD TYPE "
492 "must give a format specifier that is the "
493 "same type as that of the CASE column "
494 "specification given on FILE TYPE."));
498 else if (lex_match_id ("DUPLICATE"))
501 if (lex_match_id ("WARN"))
502 rct->flags |= RCT_DUPLICATE;
503 else if (lex_match_id ("NOWARN"))
504 rct->flags &= ~RCT_DUPLICATE;
507 msg (SE, _("WARN or NOWARN expected on DUPLICATE "
512 else if (lex_match_id ("MISSING"))
515 if (lex_match_id ("WARN"))
516 rct->flags |= RCT_MISSING;
517 else if (lex_match_id ("NOWARN"))
518 rct->flags &= ~RCT_MISSING;
521 msg (SE, _("WARN or NOWARN expected on MISSING subcommand."));
525 else if (lex_match_id ("SPREAD"))
528 if (lex_match_id ("YES"))
529 rct->flags |= RCT_SPREAD;
530 else if (lex_match_id ("NO"))
531 rct->flags &= ~RCT_SPREAD;
534 msg (SE, _("YES or NO expected on SPREAD subcommand."));
540 lex_error (_("while expecting a valid subcommand"));
546 fty->recs_tail = fty->recs_tail->next = xmalloc (sizeof *fty->recs_tail);
548 fty->recs_head = fty->recs_tail = xmalloc (sizeof *fty->recs_tail);
549 memcpy (fty->recs_tail, &rct, sizeof *fty->recs_tail);
554 if (formats[fty->record.fmt].cat & FCAT_STRING)
558 for (i = 0; i < rct->nv; i++)
570 cmd_end_file_type (void)
572 struct file_type_pgm *fty;
574 if (pgm_state != STATE_INPUT
575 || case_source_is_class (vfm_source, &file_type_source_class))
577 msg (SE, _("This command may only appear within a "
578 "FILE TYPE/END FILE TYPE structure."));
581 fty = vfm_source->aux;
583 lex_match_id ("TYPE");
587 fty->recs_tail->lt = n_trns - 1;
588 if (!(fty->recs_tail->flags & RCT_SKIP)
589 && fty->recs_tail->ft == fty->recs_tail->lt)
591 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
592 "on above RECORD TYPE."));
598 msg (SE, _("No commands between FILE TYPE and END FILE TYPE."));
604 return lex_end_of_command ();
607 /* Come here on discovering catastrophic error. */
609 discard_variables ();
613 /* FILE TYPE runtime. */
615 /*static void read_from_file_type_mixed(void);
616 static void read_from_file_type_grouped(void);
617 static void read_from_file_type_nested(void); */
619 /* Reads any number of cases into temp_case and calls write_case() for
620 each one. Compare data-list.c:read_from_data_list. */
622 file_type_source_read (struct case_source *source,
623 write_case_func *write_case UNUSED,
624 write_case_data wc_data UNUSED)
626 struct file_type_pgm *fty = source->aux;
630 struct fmt_spec format;
632 dfm_push (fty->handle);
634 format.type = fty->record.fmt;
635 format.w = fty->record.nc;
637 while (NULL != (line = dfm_get_record (fty->handle, &len)))
639 struct record_type *iter;
643 if (formats[fty->record.fmt].cat & FCAT_STRING)
647 v.c = temp_case->data[fty->record.v->fv].s;
649 data_in_finite_line (&di, line, len,
650 fty->record.fc, fty->record.fc + fty->record.nc);
651 di.v = (union value *) v.c;
653 di.f1 = fty->record.fc;
657 for (iter = fty->recs_head; iter; iter = iter->next)
659 if (iter->flags & RCT_OTHER)
661 for (i = 0; i < iter->nv; i++)
662 if (!memcmp (iter->v[i].c, v.c, fty->record.nc))
666 msg (SW, _("Unknown record type \"%.*s\"."), fty->record.nc, v.c);
672 data_in_finite_line (&di, line, len,
673 fty->record.fc, fty->record.fc + fty->record.nc);
676 di.f1 = fty->record.fc;
680 memcpy (&temp_case->data[fty->record.v->fv].f, &v.f, sizeof v.f);
681 for (iter = fty->recs_head; iter; iter = iter->next)
683 if (iter->flags & RCT_OTHER)
685 for (i = 0; i < iter->nv; i++)
686 if (iter->v[i].f == v.f)
690 msg (SW, _("Unknown record type %g."), v.f);
692 dfm_fwd_record (fty->handle);
696 /* Arrive here if there is a matching record_type, which is in
698 dfm_fwd_record (fty->handle);
703 case FTY_MIXED: read_from_file_type_mixed(); break;
704 case FTY_GROUPED: read_from_file_type_grouped(); break;
705 case FTY_NESTED: read_from_file_type_nested(); break;
709 dfm_pop (fty->handle);
713 file_type_source_destroy (struct case_source *source)
715 struct file_type_pgm *fty = source->aux;
716 struct record_type *iter, *next;
718 cancel_transformations ();
719 for (iter = fty->recs_head; iter; iter = next)
726 const struct case_source_class file_type_source_class =
729 file_type_source_read,
730 file_type_source_destroy,