1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2017 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
19 #include "data/case.h"
20 #include "data/casereader.h"
21 #include "data/casewriter.h"
22 #include "data/dataset.h"
23 #include "data/dictionary.h"
24 #include "data/format.h"
25 #include "data/transformations.h"
26 #include "data/variable.h"
27 #include "language/command.h"
28 #include "language/data-io/data-parser.h"
29 #include "language/data-io/data-reader.h"
30 #include "language/data-io/file-handle.h"
31 #include "language/data-io/inpt-pgm.h"
32 #include "language/data-io/placement-parser.h"
33 #include "language/lexer/lexer.h"
34 #include "language/lexer/variable-parser.h"
35 #include "libpspp/i18n.h"
36 #include "libpspp/message.h"
37 #include "libpspp/misc.h"
40 #include "gl/xalloc.h"
43 #define _(msgid) gettext (msgid)
45 /* DATA LIST transformation data. */
48 struct data_parser *parser; /* Parser. */
49 struct dfm_reader *reader; /* Data file reader. */
50 struct variable *end; /* Variable specified on END subcommand. */
53 static trns_free_func data_list_trns_free;
54 static trns_proc_func data_list_trns_proc;
69 static const int ROWTYPE_WIDTH = 8;
73 enum triangle triangle;
74 enum diagonal diagonal;
75 const struct variable *rowtype;
76 const struct variable *varname;
77 int n_continuous_vars;
78 struct variable **split_vars;
84 valid rowtype_ values:
101 /* Sets the value of OUTCASE which corresponds to VNAME
102 to the value STR. VNAME must be of type string.
105 set_varname_column (struct ccase *outcase, const struct variable *vname,
108 int len = var_get_width (vname);
109 uint8_t *s = value_str_rw (case_data_rw (outcase, vname), len);
111 strncpy ((char *) s, str, len);
115 blank_varname_column (struct ccase *outcase, const struct variable *vname)
117 int len = var_get_width (vname);
118 uint8_t *s = value_str_rw (case_data_rw (outcase, vname), len);
120 memset (s, ' ', len);
123 static struct casereader *
124 preprocess (struct casereader *casereader0, const struct dictionary *dict, void *aux)
126 struct matrix_format *mformat = aux;
127 const struct caseproto *proto = casereader_get_proto (casereader0);
128 struct casewriter *writer;
129 writer = autopaging_writer_create (proto);
130 struct ccase *prev_case = NULL;
131 double **matrices = NULL;
134 const size_t sizeof_matrix =
135 sizeof (double) * mformat->n_continuous_vars * mformat->n_continuous_vars;
138 /* Make an initial pass to populate our temporary matrix */
139 struct casereader *pass0 = casereader_clone (casereader0);
141 unsigned int prev_split_hash = 1;
142 int row = (mformat->triangle == LOWER && mformat->diagonal == NO_DIAGONAL) ? 1 : 0;
143 for (; (c = casereader_read (pass0)) != NULL; case_unref (c))
146 unsigned int split_hash = 0;
147 for (s = 0; s < mformat->n_split_vars; ++s)
149 const struct variable *svar = mformat->split_vars[s];
150 const union value *sv = case_data (c, svar);
151 split_hash = value_hash (sv, var_get_width (svar), split_hash);
154 if (matrices == NULL || prev_split_hash != split_hash)
156 row = (mformat->triangle == LOWER && mformat->diagonal == NO_DIAGONAL) ?
160 matrices = xrealloc (matrices, sizeof (double*) * n_splits);
161 matrices[n_splits - 1] = xmalloc (sizeof_matrix);
164 prev_split_hash = split_hash;
166 int c_offset = (mformat->triangle == UPPER) ? row : 0;
167 if (mformat->triangle == UPPER && mformat->diagonal == NO_DIAGONAL)
169 const union value *v = case_data (c, mformat->rowtype);
170 const char *val = (const char *) value_str (v, ROWTYPE_WIDTH);
171 if (0 == strncasecmp (val, "corr ", ROWTYPE_WIDTH) ||
172 0 == strncasecmp (val, "cov ", ROWTYPE_WIDTH))
174 if (row >= mformat->n_continuous_vars)
177 _("There are %d variable declared but the data has at least %d matrix rows."),
178 mformat->n_continuous_vars, row + 1);
182 for (col = c_offset; col < mformat->n_continuous_vars; ++col)
184 const struct variable *var =
187 var_get_dict_index (mformat->varname));
189 double e = case_data (c, var)->f;
193 /* Fill in the lower triangle */
194 (matrices[n_splits-1])[col + mformat->n_continuous_vars * row] = e;
196 if (mformat->triangle != FULL)
197 /* Fill in the upper triangle */
198 (matrices[n_splits-1]) [row + mformat->n_continuous_vars * col] = e;
203 casereader_destroy (pass0);
205 /* Now make a second pass to fill in the other triangle from our
207 const int idx = var_get_dict_index (mformat->varname);
213 struct ccase *outcase = case_create (proto);
214 union value *v = case_data_rw (outcase, mformat->rowtype);
215 uint8_t *n = value_str_rw (v, ROWTYPE_WIDTH);
216 strncpy ((char *) n, "N ", ROWTYPE_WIDTH);
217 blank_varname_column (outcase, mformat->varname);
218 for (col = 0; col < mformat->n_continuous_vars; ++col)
220 union value *dest_val =
221 case_data_rw_idx (outcase,
222 1 + col + var_get_dict_index (mformat->varname));
223 dest_val->f = mformat->n;
225 casewriter_write (writer, outcase);
230 for (; (c = casereader_read (casereader0)) != NULL; prev_case = c)
233 unsigned int split_hash = 0;
234 for (s = 0; s < mformat->n_split_vars; ++s)
236 const struct variable *svar = mformat->split_vars[s];
237 const union value *sv = case_data (c, svar);
238 split_hash = value_hash (sv, var_get_width (svar), split_hash);
240 if (prev_split_hash != split_hash)
246 prev_split_hash = split_hash;
247 case_unref (prev_case);
248 const union value *v = case_data (c, mformat->rowtype);
249 const char *val = (const char *) value_str (v, ROWTYPE_WIDTH);
252 if (0 == strncasecmp (val, "n ", ROWTYPE_WIDTH) ||
253 0 == strncasecmp (val, "n_vector", ROWTYPE_WIDTH))
256 _("The N subcommand was specified, but a N record was also found in the data. The N record will be ignored."));
261 struct ccase *outcase = case_create (proto);
262 case_copy (outcase, 0, c, 0, caseproto_get_n_widths (proto));
264 if (0 == strncasecmp (val, "corr ", ROWTYPE_WIDTH) ||
265 0 == strncasecmp (val, "cov ", ROWTYPE_WIDTH))
268 const struct variable *var = dict_get_var (dict, idx + 1 + row);
269 set_varname_column (outcase, mformat->varname, var_get_name (var));
270 value_copy (case_data_rw (outcase, mformat->rowtype), v, ROWTYPE_WIDTH);
272 for (col = 0; col < mformat->n_continuous_vars; ++col)
274 union value *dest_val =
275 case_data_rw_idx (outcase,
276 1 + col + var_get_dict_index (mformat->varname));
277 dest_val->f = (matrices[n_splits - 1])[col + mformat->n_continuous_vars * row];
278 if (col == row && mformat->diagonal == NO_DIAGONAL)
285 blank_varname_column (outcase, mformat->varname);
288 /* Special case for SD and N_VECTOR: Rewrite as STDDEV and N respectively */
289 if (0 == strncasecmp (val, "sd ", ROWTYPE_WIDTH))
291 value_copy_buf_rpad (case_data_rw (outcase, mformat->rowtype), ROWTYPE_WIDTH,
292 (uint8_t *) "STDDEV", 6, ' ');
294 else if (0 == strncasecmp (val, "n_vector", ROWTYPE_WIDTH))
296 value_copy_buf_rpad (case_data_rw (outcase, mformat->rowtype), ROWTYPE_WIDTH,
297 (uint8_t *) "N", 1, ' ');
300 casewriter_write (writer, outcase);
303 /* If NODIAGONAL is specified, then a final case must be written */
304 if (mformat->diagonal == NO_DIAGONAL)
307 struct ccase *outcase = case_create (proto);
310 case_copy (outcase, 0, prev_case, 0, caseproto_get_n_widths (proto));
312 const struct variable *var = dict_get_var (dict, idx + 1 + row);
313 set_varname_column (outcase, mformat->varname, var_get_name (var));
315 for (col = 0; col < mformat->n_continuous_vars; ++col)
317 union value *dest_val =
318 case_data_rw_idx (outcase, 1 + col +
319 var_get_dict_index (mformat->varname));
320 dest_val->f = (matrices[n_splits - 1]) [col + mformat->n_continuous_vars * row];
321 if (col == row && mformat->diagonal == NO_DIAGONAL)
325 casewriter_write (writer, outcase);
330 case_unref (prev_case);
333 for (i = 0 ; i < n_splits; ++i)
336 struct casereader *reader1 = casewriter_make_reader (writer);
337 casereader_destroy (casereader0);
343 case_unref (prev_case);
345 for (i = 0 ; i < n_splits; ++i)
348 casereader_destroy (casereader0);
353 cmd_matrix (struct lexer *lexer, struct dataset *ds)
355 struct dictionary *dict;
356 struct data_parser *parser;
357 struct dfm_reader *reader;
358 struct file_handle *fh = NULL;
359 char *encoding = NULL;
360 struct matrix_format mformat;
365 mformat.triangle = LOWER;
366 mformat.diagonal = DIAGONAL;
367 mformat.n_split_vars = 0;
368 mformat.split_vars = NULL;
371 dict = (in_input_program ()
373 : dict_create (get_default_encoding ()));
374 parser = data_parser_create (dict);
377 data_parser_set_type (parser, DP_DELIMITED);
378 data_parser_set_warn_missing_fields (parser, false);
379 data_parser_set_span (parser, false);
381 mformat.rowtype = dict_create_var (dict, "ROWTYPE_", ROWTYPE_WIDTH);
383 mformat.n_continuous_vars = 0;
384 mformat.n_split_vars = 0;
386 if (! lex_force_match_id (lexer, "VARIABLES"))
389 lex_match (lexer, T_EQUALS);
391 if (! parse_mixed_vars (lexer, dict, &names, &n_names, PV_NO_DUPLICATE))
394 for (i = 0; i < n_names; ++i)
400 int longest_name = 0;
401 for (i = 0; i < n_names; ++i)
403 maximize_int (&longest_name, strlen (names[i]));
406 mformat.varname = dict_create_var (dict, "VARNAME_",
407 8 * DIV_RND_UP (longest_name, 8));
409 for (i = 0; i < n_names; ++i)
411 if (0 == strcasecmp (names[i], "ROWTYPE_"))
413 const struct fmt_spec fmt = fmt_for_input (FMT_A, 8, 0);
414 data_parser_add_delimited_field (parser,
416 var_get_case_index (mformat.rowtype),
421 const struct fmt_spec fmt = fmt_for_input (FMT_F, 10, 4);
422 struct variable *v = dict_create_var (dict, names[i], 0);
423 var_set_both_formats (v, &fmt);
424 data_parser_add_delimited_field (parser,
426 var_get_case_index (mformat.varname) +
427 ++mformat.n_continuous_vars,
431 for (i = 0; i < n_names; ++i)
435 while (lex_token (lexer) != T_ENDCMD)
437 if (! lex_force_match (lexer, T_SLASH))
440 if (lex_match_id (lexer, "N"))
442 lex_match (lexer, T_EQUALS);
444 if (! lex_force_int (lexer))
447 mformat.n = lex_integer (lexer);
450 msg (SE, _("%s must not be negative."), "N");
455 else if (lex_match_id (lexer, "FORMAT"))
457 lex_match (lexer, T_EQUALS);
459 while (lex_token (lexer) != T_SLASH && (lex_token (lexer) != T_ENDCMD))
461 if (lex_match_id (lexer, "LIST"))
463 data_parser_set_span (parser, false);
465 else if (lex_match_id (lexer, "FREE"))
467 data_parser_set_span (parser, true);
469 else if (lex_match_id (lexer, "UPPER"))
471 mformat.triangle = UPPER;
473 else if (lex_match_id (lexer, "LOWER"))
475 mformat.triangle = LOWER;
477 else if (lex_match_id (lexer, "FULL"))
479 mformat.triangle = FULL;
481 else if (lex_match_id (lexer, "DIAGONAL"))
483 mformat.diagonal = DIAGONAL;
485 else if (lex_match_id (lexer, "NODIAGONAL"))
487 mformat.diagonal = NO_DIAGONAL;
491 lex_error (lexer, NULL);
496 else if (lex_match_id (lexer, "FILE"))
498 lex_match (lexer, T_EQUALS);
500 fh = fh_parse (lexer, FH_REF_FILE | FH_REF_INLINE, NULL);
504 else if (lex_match_id (lexer, "SPLIT"))
506 lex_match (lexer, T_EQUALS);
507 if (! parse_variables (lexer, dict, &mformat.split_vars, &mformat.n_split_vars, 0))
509 free (mformat.split_vars);
513 for (i = 0; i < mformat.n_split_vars; ++i)
515 const struct fmt_spec fmt = fmt_for_input (FMT_F, 4, 0);
516 var_set_both_formats (mformat.split_vars[i], &fmt);
518 dict_reorder_vars (dict, mformat.split_vars, mformat.n_split_vars);
519 mformat.n_continuous_vars -= mformat.n_split_vars;
523 lex_error (lexer, NULL);
528 if (mformat.diagonal == NO_DIAGONAL && mformat.triangle == FULL)
530 msg (SE, _("FORMAT = FULL and FORMAT = NODIAGONAL are mutually exclusive."));
535 fh = fh_inline_file ();
536 fh_set_default_handle (fh);
538 if (!data_parser_any_fields (parser))
540 msg (SE, _("At least one variable must be specified."));
544 if (lex_end_of_command (lexer) != CMD_SUCCESS)
547 reader = dfm_open_reader (fh, lexer, encoding);
551 if (in_input_program ())
553 struct data_list_trns *trns = xmalloc (sizeof *trns);
554 trns->parser = parser;
555 trns->reader = reader;
557 add_transformation (ds, data_list_trns_proc, data_list_trns_free, trns);
561 data_parser_make_active_file (parser, ds, reader, dict, preprocess, &mformat);
566 free (mformat.split_vars);
568 return CMD_DATA_LIST;
571 data_parser_destroy (parser);
572 if (!in_input_program ())
576 free (mformat.split_vars);
577 return CMD_CASCADING_FAILURE;
581 /* Input procedure. */
583 /* Destroys DATA LIST transformation TRNS.
584 Returns true if successful, false if an I/O error occurred. */
586 data_list_trns_free (void *trns_)
588 struct data_list_trns *trns = trns_;
589 data_parser_destroy (trns->parser);
590 dfm_close_reader (trns->reader);
595 /* Handle DATA LIST transformation TRNS, parsing data into *C. */
597 data_list_trns_proc (void *trns_, struct ccase **c, casenumber case_num UNUSED)
599 struct data_list_trns *trns = trns_;
602 *c = case_unshare (*c);
603 if (data_parser_parse (trns->parser, trns->reader, *c))
604 retval = TRNS_CONTINUE;
605 else if (dfm_reader_error (trns->reader) || dfm_eof (trns->reader) > 1)
607 /* An I/O error, or encountering end of file for a second
608 time, should be escalated into a more serious error. */
612 retval = TRNS_END_FILE;
614 /* If there was an END subcommand handle it. */
615 if (trns->end != NULL)
617 double *end = &case_data_rw (*c, trns->end)->f;
618 if (retval == TRNS_END_FILE)
621 retval = TRNS_CONTINUE;