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 "dictionary.h"
29 #include "file-handle.h"
36 /* Defines the three types of complex files read by FILE TYPE. */
44 /* Limited variable column specifications. */
47 char name[9]; /* Variable name. */
48 int fc, nc; /* First column (1-based), # of columns. */
49 int fmt; /* Format type. */
50 struct variable *v; /* Variable. */
53 /* RCT_* record type constants. */
56 RCT_OTHER = 001, /* 1=OTHER. */
57 RCT_SKIP = 002, /* 1=SKIP. */
58 RCT_DUPLICATE = 004, /* DUPLICATE: 0=NOWARN, 1=WARN. */
59 RCT_MISSING = 010, /* MISSING: 0=NOWARN, 1=WARN. */
60 RCT_SPREAD = 020 /* SPREAD: 0=NO, 1=YES. */
63 /* Represents a RECORD TYPE command. */
66 struct record_type *next;
67 unsigned flags; /* RCT_* constants. */
68 union value *v; /* Vector of values for this record type. */
69 int nv; /* Length of vector V. */
70 struct col_spec case_sbc; /* CASE subcommand. */
71 int ft, lt; /* First, last transformation index. */
74 /* Represents a FILE TYPE input program. Does not contain a
75 trns_header because it's never submitted as a transformation. */
78 int type; /* One of the FTY_* constants. */
79 struct dfm_reader *reader; /* Data file to read. */
80 struct col_spec record; /* RECORD subcommand. */
81 struct col_spec case_sbc; /* CASE subcommand. */
82 int wild; /* 0=NOWARN, 1=WARN. */
83 int duplicate; /* 0=NOWARN, 1=WARN. */
84 int missing; /* 0=NOWARN, 1=WARN, 2=CASE. */
85 int ordered; /* 0=NO, 1=YES. */
86 int had_rec_type; /* 1=Had a RECORD TYPE command.
87 RECORD TYPE must precede the first
89 struct record_type *recs_head; /* List of record types. */
90 struct record_type *recs_tail; /* Last in list of record types. */
91 size_t case_size; /* Case size in bytes. */
94 static int parse_col_spec (struct col_spec *, const char *);
95 static void create_col_var (struct col_spec *c);
97 /* Parses FILE TYPE command. */
101 static struct file_type_pgm *fty; /* FIXME: static? WTF? */
102 struct file_handle *fh = NULL;
105 discard_variables ();
107 fty = xmalloc (sizeof *fty);
109 fty->record.name[0] = 0;
110 fty->case_sbc.name[0] = 0;
111 fty->wild = fty->duplicate = fty->missing = fty->ordered = 0;
112 fty->had_rec_type = 0;
113 fty->recs_head = fty->recs_tail = NULL;
115 if (lex_match_id ("MIXED"))
116 fty->type = FTY_MIXED;
117 else if (lex_match_id ("GROUPED"))
119 fty->type = FTY_GROUPED;
125 else if (lex_match_id ("NESTED"))
126 fty->type = FTY_NESTED;
129 msg (SE, _("MIXED, GROUPED, or NESTED expected."));
135 if (lex_match_id ("FILE"))
142 else if (lex_match_id ("RECORD"))
145 if (!parse_col_spec (&fty->record, "####RECD"))
148 else if (lex_match_id ("CASE"))
150 if (fty->type == FTY_MIXED)
152 msg (SE, _("The CASE subcommand is not valid on FILE TYPE "
158 if (!parse_col_spec (&fty->case_sbc, "####CASE"))
161 else if (lex_match_id ("WILD"))
164 if (lex_match_id ("WARN"))
166 else if (lex_match_id ("NOWARN"))
170 msg (SE, _("WARN or NOWARN expected after WILD."));
174 else if (lex_match_id ("DUPLICATE"))
176 if (fty->type == FTY_MIXED)
178 msg (SE, _("The DUPLICATE subcommand is not valid on "
179 "FILE TYPE MIXED."));
184 if (lex_match_id ("WARN"))
186 else if (lex_match_id ("NOWARN"))
188 else if (lex_match_id ("CASE"))
190 if (fty->type != FTY_NESTED)
192 msg (SE, _("DUPLICATE=CASE is only valid on "
193 "FILE TYPE NESTED."));
201 msg (SE, _("WARN%s expected after DUPLICATE."),
202 (fty->type == FTY_NESTED ? _(", NOWARN, or CASE")
207 else if (lex_match_id ("MISSING"))
209 if (fty->type == FTY_MIXED)
211 msg (SE, _("The MISSING subcommand is not valid on "
212 "FILE TYPE MIXED."));
217 if (lex_match_id ("NOWARN"))
219 else if (lex_match_id ("WARN"))
223 msg (SE, _("WARN or NOWARN after MISSING."));
227 else if (lex_match_id ("ORDERED"))
229 if (fty->type != FTY_GROUPED)
231 msg (SE, _("ORDERED is only valid on FILE TYPE GROUPED."));
236 if (lex_match_id ("YES"))
238 else if (lex_match_id ("NO"))
242 msg (SE, _("YES or NO expected after ORDERED."));
248 lex_error (_("while expecting a valid subcommand"));
253 if (fty->record.name[0] == 0)
255 msg (SE, _("The required RECORD subcommand was not present."));
259 if (fty->type == FTY_GROUPED)
261 if (fty->case_sbc.name[0] == 0)
263 msg (SE, _("The required CASE subcommand was not present."));
267 if (!strcmp (fty->case_sbc.name, fty->record.name))
269 msg (SE, _("CASE and RECORD must specify different variable "
275 fty->reader = dfm_open_reader (fh);
276 if (fty->reader == NULL)
280 create_col_var (&fty->record);
281 if (fty->case_sbc.name[0])
282 create_col_var (&fty->case_sbc);
283 vfm_source = create_case_source (&file_type_source_class, default_dict, fty);
292 /* Creates a variable with attributes specified by struct col_spec C, and
293 stores it into C->V. */
295 create_col_var (struct col_spec *c)
299 if (formats[c->fmt].cat & FCAT_STRING)
303 c->v = dict_create_var (default_dict, c->name, width);
306 /* Parses variable, column, type specifications for a variable. */
308 parse_col_spec (struct col_spec *c, const char *def_name)
310 struct fmt_spec spec;
315 strcpy (c->name, tokid);
319 strcpy (c->name, def_name);
322 if (!lex_force_int ())
324 c->fc = lex_integer ();
327 msg (SE, _("Column value must be positive."));
333 lex_negative_to_dash ();
336 if (!lex_force_int ())
338 c->nc = lex_integer ();
343 msg (SE, _("Ending column precedes beginning column."));
352 /* Format specifier. */
356 if (!lex_force_id ())
358 c->fmt = parse_format_specifier_name (&cp, 0);
363 msg (SE, _("Bad format specifier name."));
367 if (!lex_force_match (')'))
376 return check_input_specifier (&spec);
381 /* Parse the RECORD TYPE command. */
383 cmd_record_type (void)
385 struct file_type_pgm *fty;
386 struct record_type *rct;
388 /* Make sure we're inside a FILE TYPE structure. */
389 if (pgm_state != STATE_INPUT
390 || !case_source_is_class (vfm_source, &file_type_source_class))
392 msg (SE, _("This command may only appear within a "
393 "FILE TYPE/END FILE TYPE structure."));
397 fty = vfm_source->aux;
399 /* Initialize the record_type structure. */
400 rct = xmalloc (sizeof *rct);
404 rct->flags |= RCT_DUPLICATE;
406 rct->flags |= RCT_MISSING;
410 if (fty->case_sbc.name[0])
411 rct->case_sbc = fty->case_sbc;
413 if (fty->recs_tail && (fty->recs_tail->flags & RCT_OTHER))
415 msg (SE, _("OTHER may appear only on the last RECORD TYPE command."));
421 fty->recs_tail->lt = n_trns - 1;
422 if (!(fty->recs_tail->flags & RCT_SKIP)
423 && fty->recs_tail->ft == fty->recs_tail->lt)
425 msg (SE, _("No input commands (DATA LIST, REPEATING DATA) "
426 "for above RECORD TYPE."));
431 /* Parse record type values. */
432 if (lex_match_id ("OTHER"))
433 rct->flags |= RCT_OTHER;
438 while (token == T_NUM || token == T_STRING)
443 rct->v = xrealloc (rct->v, mv * sizeof *rct->v);
446 if (formats[fty->record.fmt].cat & FCAT_STRING)
448 if (!lex_force_string ())
450 rct->v[rct->nv].c = xmalloc (fty->record.nc + 1);
451 st_bare_pad_copy (rct->v[rct->nv].c, ds_c_str (&tokstr),
456 if (!lex_force_num ())
458 rct->v[rct->nv].f = tokval;
467 /* Parse the rest of the subcommands. */
470 if (lex_match_id ("SKIP"))
471 rct->flags |= RCT_SKIP;
472 else if (lex_match_id ("CASE"))
474 if (fty->type == FTY_MIXED)
476 msg (SE, _("The CASE subcommand is not allowed on "
477 "the RECORD TYPE command for FILE TYPE MIXED."));
482 if (!parse_col_spec (&rct->case_sbc, ""))
484 if (rct->case_sbc.name[0])
486 msg (SE, _("No variable name may be specified for the "
487 "CASE subcommand on RECORD TYPE."));
491 if ((formats[rct->case_sbc.fmt].cat ^ formats[fty->case_sbc.fmt].cat)
494 msg (SE, _("The CASE column specification on RECORD TYPE "
495 "must give a format specifier that is the "
496 "same type as that of the CASE column "
497 "specification given on FILE TYPE."));
501 else if (lex_match_id ("DUPLICATE"))
504 if (lex_match_id ("WARN"))
505 rct->flags |= RCT_DUPLICATE;
506 else if (lex_match_id ("NOWARN"))
507 rct->flags &= ~RCT_DUPLICATE;
510 msg (SE, _("WARN or NOWARN expected on DUPLICATE "
515 else if (lex_match_id ("MISSING"))
518 if (lex_match_id ("WARN"))
519 rct->flags |= RCT_MISSING;
520 else if (lex_match_id ("NOWARN"))
521 rct->flags &= ~RCT_MISSING;
524 msg (SE, _("WARN or NOWARN expected on MISSING subcommand."));
528 else if (lex_match_id ("SPREAD"))
531 if (lex_match_id ("YES"))
532 rct->flags |= RCT_SPREAD;
533 else if (lex_match_id ("NO"))
534 rct->flags &= ~RCT_SPREAD;
537 msg (SE, _("YES or NO expected on SPREAD subcommand."));
543 lex_error (_("while expecting a valid subcommand"));
549 fty->recs_tail = fty->recs_tail->next = xmalloc (sizeof *fty->recs_tail);
551 fty->recs_head = fty->recs_tail = xmalloc (sizeof *fty->recs_tail);
552 memcpy (fty->recs_tail, &rct, sizeof *fty->recs_tail);
557 if (formats[fty->record.fmt].cat & FCAT_STRING)
561 for (i = 0; i < rct->nv; i++)
573 cmd_end_file_type (void)
575 struct file_type_pgm *fty;
577 if (pgm_state != STATE_INPUT
578 || case_source_is_class (vfm_source, &file_type_source_class))
580 msg (SE, _("This command may only appear within a "
581 "FILE TYPE/END FILE TYPE structure."));
584 fty = vfm_source->aux;
585 fty->case_size = dict_get_case_size (default_dict);
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;
630 struct fmt_spec format;
632 dfm_push (fty->reader);
634 format.type = fty->record.fmt;
635 format.w = fty->record.nc;
637 while (!dfm_eof (fty->reader))
639 struct len_string line;
640 struct record_type *iter;
644 dfm_expand_tabs (fty->reader);
645 dfm_get_record (fty->reader, &line);
646 if (formats[fty->record.fmt].cat & FCAT_STRING)
650 v.c = case_data_rw (c, fty->record.v->fv)->s;
652 data_in_finite_line (&di, ls_c_str (&line), ls_length (&line),
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, ls_c_str (&line), ls_length (&line),
676 fty->record.fc, fty->record.fc + fty->record.nc);
679 di.f1 = fty->record.fc;
683 case_data_rw (c, fty->record.v->fv)->f = 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_forward_record (fty->reader);
699 /* Arrive here if there is a matching record_type, which is in
701 dfm_forward_record (fty->reader);
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->reader);
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 dfm_close_reader (fty->reader);
723 for (iter = fty->recs_head; iter; iter = next)
730 const struct case_source_class file_type_source_class =
734 file_type_source_read,
735 file_type_source_destroy,