1 2008-01-02 John Darrington <john@darrington.wattle.id.au>
3 * binomial.c chisquare.c examine.q frequencies.q oneway.q regression.q : updated
4 all users of var_get_value_name to use replacement function var_append_value_name.
6 2007-12-07 Ben Pfaff <blp@gnu.org>
10 * crosstabs.q (precalc): Initialize data structures even if the
11 first case cannot be read.
13 * frequencies.q (precalc): Ditto.
15 2007-11-03 Ben Pfaff <blp@gnu.org>
17 Allow output files to overwrite input files (bug #21280).
19 * aggregate.c (cmd_aggregate): Manage file handle reference
22 * correlations.q (internal_cmd_frequencies): Ditto.
23 (cor_custom_matrix): Ditto.
25 * regression.q (regression_custom_export): Ditto.
26 (cmd_regression): Ditto.
28 2007-10-12 Ben Pfaff <blp@gnu.org>
30 * flip.c (flip_file): No need to conditionally substitute for
31 "fseeko" and "off_t" manually anymore, as gnulib takes care of it
34 2007-09-21 Jason Stover <jhs@wonko.gcsu.edu>
36 * regression.q (run_regression): Partial fix of memory leak, bug
39 2007-09-19 Ben Pfaff <blp@gnu.org>
43 * aggregate.c (cmd_aggregate): Destroy casereader consistently,
44 even if casereader fails.
46 * examine.q (run_examine): Ditto.
48 * glm.q (run_glm): Ditto.
50 * oneway.q (run_oneway): Ditto.
52 * regression.q (run_regression): Ditto.
54 * t-test.q (calculate): Ditto.
56 * descriptives.c (calc_descriptives): Ditto. Also avoid
57 gratuitous casereader_clone.
59 2007-09-13 Jason Stover <jhs@debs.hoobahooba.net>
61 * regression.q (cmd_regression): Move declaration of models in to
62 definition of cmd_regression.
64 * regression.q (run_regression): Free mom to fix memory leak.
66 2007-09-12 Ben Pfaff <blp@gnu.org>
68 * crosstabs.q (postcalc): Free sorted_tab and the structures that
69 it points to, to plug a memory leak. Fixes bug #20910. Thanks to
70 John Darrington for reporting this bug and for review.
72 2007-09-04 Ben Pfaff <blp@gnu.org>
74 * crosstabs.q (cmd_crosstabs): Free xtab and the structures that
75 it points to, to plug a memory leak. Fixes bug #18315.
77 2007-08-15 Jason Stover <jhs@wonko.gcsu.edu>
79 * regression.q (identify_indep_vars): Print an error if dependent
80 and independent variables are the same. Fixes bug 19819.
82 2007-08-12 Ben Pfaff <blp@gnu.org>
84 * flip.c: Drop use of dict_get_compacted_dict_index_to_case_index
85 and just use the ordinary case indexes. There seemed to be no
86 reason for the former method.
88 2007-08-03 Ben Pfaff <blp@gnu.org>
90 * rank.q (rank_cmd): Instead of sorting by SPLIT FILE vars, group
91 by them. Fixes bug #17239.
92 Reviewed by John Darrington.
94 2007-08-01 Ben Pfaff <blp@gnu.org>
96 Clean up handling of median, by treating it almost like any other
97 percentile. Fixes bug #17424. Thanks to John Darrington for
99 * frequencies.q (internal_cmd_frequencies): Fix handling of bit
100 masks for `stats' variable. If median is selected, turn it off
101 and add a 50th percentile.
102 (add_percentile): Simplify code a little.
103 (calc_stats): Drop special casing of median.
104 (dump_statistics): Ditto, except that we label the 50th percentile
105 as "50 (Median)" to make it clear that it's also the median.
107 2007-07-31 Ben Pfaff <blp@gnu.org>
109 Remove integer mode from FREQUENCIES and incidentally fix bug
110 #17421. Reviewed by John Darrington.
111 * frequencies.q (int_pool): Rename data_pool.
112 (gen_pool): Rename syntax_pool.
113 (enum FRQM_*): Removed.
114 (struct freq_tab): Removed `mode', `vector', `min', `max',
115 `out_of_range', `sysmis' members.
116 (calc): Delete support for integer mode.
118 (postprocess_freq_tab): Ditto.
119 (cleanup_freq_tab): Ditto.
120 (frq_custom_variables): Ditto.
122 2007-07-28 John Darrington <john@darrington.wattle.id.au>
124 * t-test.q: Moved the order in which groups are displayed in the
125 independent samples case, where a cut point is given.
127 2007-07-27 Ben Pfaff <blp@gnu.org>
129 * regression.q (run_regression): Move casereader_destroy call so
130 that it always gets called, not just if there was some valid
131 data. Fixes bug #19581.
132 Reviewed by Jason Stover.
134 2007-07-24 Ben Pfaff <blp@gnu.org>
136 * flip.c (struct flip_pgm): Remove `case_size' member (now
138 (cmd_flip): Pass var_cnt as number of cases instead of case_cnt,
139 to fix bug #20494. Don't assign to `case_size' member, which was
140 unused after assignment.
141 (build_dictionary): When NEWNAMES not used, get the number of
142 variables right, to fix bug #20493.
144 2007-07-10 Jason Stover <jhs@math.gcsu.edu>
146 * glm.q: Initial version of the GLM procedure.
148 2007-06-06 Ben Pfaff <blp@gnu.org>
150 Adapt case sources, sinks, and clients of procedure code to the
153 * aggregate.c: Simplify greatly since everything is more uniform
156 * autorecode.c: Adapt to new procedure code.
158 * chisquare.c: Ditto.
159 * crosstabs.q: Ditto.
160 * descriptives.c: Ditto.
162 * npar-summary.c: Ditto.
163 * frequencies.q: Ditto.
166 * regression.q: Ditto.
167 * sort-cases.c: Ditto.
170 * sort-criteria.c: Rewrite to output a struct case_ordering.
172 * flip.c: Rewrite to be a casereader.
174 * rank.q: Simplify greatly since casereaders are much more
175 flexible than what we had before.
177 2007-05-15 Jason Stover <jhs@math.gcsu.edu>
179 * regression.q (run_regression): Tell the user when the data
180 contain no valid cases.
182 2007-05-08 Jason Stover <jhs@math.gcsu.edu>
184 * regression.q: Partial fix of bug which caused a crash if
185 dependent variable and independent variable were the same.
187 2007-04-16 John Darrington <john@darrington.wattle.id.au>
189 * t-test.q: Changed the output width of reported counts and
190 degrees of freedom, to avoid truncating these values. Thanks
191 to Seth Woolley for reporting this problem. A proper fix involves
192 re-thinking the output driver.
194 2007-04-12 Jason Stover <jhs@math.gcsu.edu>
196 * regression.q (run_regression): Added if (n_data >0) to fix bug
199 2007-03-29 Jason Stover <jhs@math.gcsu.edu>
201 * regression.q (prepare_data): New function.
203 * regression.q (compute_moments): New function.
205 2007-03-18 Ben Pfaff <blp@gnu.org>
207 * crosstabs.q (static var write): Rename write_style to avoid
208 conflict with POSIX function of same name.
210 2007-03-16 Jason Stover <jhs@math.gcsu.edu>
212 * regression.q (run_regression): Added support for moments.
214 Sat Feb 17 08:16:00 2007 Ben Pfaff <blp@gnu.org>
216 * flip.c (flip_sink_create): Improve error message when temporary
217 file cannot be created.
219 Tue Feb 6 19:58:03 2007 Ben Pfaff <blp@gnu.org>
221 * flip.c (flip_file): Give better error message on end-of-file.
223 2007-02-04 Jason Stover <jhs@math.gcsu.edu>
225 * regression.q: Fixed p-value computation in the test for
226 individual regression coefficients.
228 Mon Jan 15 11:03:20 2007 Ben Pfaff <blp@gnu.org>
230 Fix bugs found by valgrind when --enable-debug is used with the
231 new case code. These bugs are hidden when the data set is small
232 enough to find in memory; if a bigger data set that would overflow
233 to disk were used, then data corruption would occur.
235 * chisquare.c (create_freq_hash): Pass free_freq_mutable_hash to
236 hsh_create as free function. Make copy of data put into hash.
238 * oneway.q (free_value): New function.
239 (run_oneway): Use free_value as arg to hsh_create. Make copy of
242 * rank.q (rank_cases): Don't access data in case after we've given
245 Tue Jan 9 19:16:11 2007 Ben Pfaff <blp@gnu.org>
249 * aggregate.c (parse_aggregate_functions) Initialize
252 Fri Dec 22 14:04:09 2006 Ben Pfaff <blp@gnu.org>
254 Simplify missing value handling.
256 * aggregate.c (struct agr_var): Remove `bool include_missing', add
257 `enum mv_class exclude'. Remove `int missing', add `bool
258 saw_missing'. Update users.
260 * descriptives.c (struct dsc_trns): Removed `int
261 include_user_missing', add `enum mv_class exclude'. Update users.
262 (struct dsc_proc): Ditto.
264 * examine.q: (static var value_is_missing): Rename
265 `exclude_values', change type to `enum mv_class'. Update users.
269 Fri Dec 22 19:22:18 WST 2006 John Darrington <john@darrington.wattle.id.au>
271 * frequencies.q : Fixed bug #17420, where the table bounds were overun
272 when /FORMAT=nolabels was given.
274 Wed Dec 20 18:45:31 WST 2006 John Darrington <john@darrington.wattle.id.au>
276 * binomial.c binomial.h : New files. Thanks to Jason Stover
277 for assistance with these.
279 * chisquare.c chisquare.h freq.c freq.h npar-summary.c npar-summary.h
280 npar.h npar.q: New files. Implementing NPAR TESTS.
282 * frequencies.q : Moved structure definitions into freq.[ch]
284 Sat Dec 16 22:26:44 2006 Ben Pfaff <blp@gnu.org>
286 Make it possible to pull cases from the active file with a
287 function call, instead of requiring indirection through a callback
290 * aggregate.c (cmd_aggregate): Take advantage of new procedure
292 (agr_to_active_file): Removed.
293 (presorted_agr_to_sysfile): Removed.
295 * autorecode.c (cmd_autorecode): Take advantage of new procedure
297 (autorecode_proc_func): Removed.
299 * flip.c (struct flip_pgm): New members to allow conformance with
300 new case_source_class interface.
301 (cmd_flip): Adapt to new case_source_class interface.
302 (flip_source_read): Ditto.
303 (flip_source_destroy): Ditto.
305 Sat Dec 16 12:54:27 2006 Ben Pfaff <blp@gnu.org>
307 * rank.q (rank_custom_variables): Allow grouping variables to be
308 strings. Fixes bug #18533. Thanks to John Darrington for review.
310 Sat Dec 9 18:47:51 2006 Ben Pfaff <blp@gnu.org>
312 * regression.q (is_depvar): Compare variable pointers instead of
315 Thu Dec 7 15:26:25 WST 2006 John Darrington <john@darrington.wattle.id.au>
317 * examine.q: Allocated the categorical values for the dependent and
318 independent variables, on the heap. Hence they can be of any width.
320 Wed Dec 6 21:14:26 2006 Ben Pfaff <blp@gnu.org>
322 * regression.q (reg_inserted): Compare variable pointers instead
325 Mon Dec 4 22:33:46 2006 Ben Pfaff <blp@gnu.org>
327 * crosstabs.q (insert_summary): Use var_to_string for labeling.
328 (output_pivot_table) Ditto.
331 * frequencies.q (setup_z_trns): Ditto.
333 (dump_condensed) Ditto.
334 (dump_statistics) Ditto.
337 Sun Nov 5 08:31:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
339 * t-test.q, oneway.q: Changed to use the new casefilter structure.
341 Sat Oct 14 16:52:28 2006 Ben Pfaff <blp@gnu.org>
343 * rank.q: (rank_sorted_casefile) Add some missing case_destroy()
344 calls to fix a memory leak.
346 Sun Oct 8 09:45:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
348 * rank.q: Plugged a small memory leak which occurred under error
351 Sat Oct 7 11:06:01 WST 2006 John Darrington <john@darrington.wattle.id.au>
353 * rank.q: Implemented most of the RANK command.
355 2006-07-14 Jason Stover <jhs@math.gcsu.edu>
357 * regression.q (run_regression): New function to move knowledge of
358 pspp_linreg_cache out of math/coefficient.[ch].
360 Sat Jul 1 17:41:46 2006 Ben Pfaff <blp@gnu.org>
362 Fix bug #11612, "q2c documentation does not agree with code".
364 * examine.q: Audit use of q2c "+" prefixes that indicate that a
365 command may appear multiple times.
367 * frequencies.q: Ditto.
371 * regression.q: Ditto.
375 Fri Jun 23 14:18:22 2006 Ben Pfaff <blp@gnu.org>
377 Support long string variables on FREQUENCIES, as
378 an extension when in enhanced algorithms mode. For Greg Hunt
379 <greg@firmansyah.com>.
381 * frequencies.q: (struct freq) Change `v' member from union value
382 to union value *. Update all references.
383 (struct var_freqs) Add width, print members to represent effective
384 variable width and display format.
385 (calc) Copy entire long string value into the hash table.
386 (frq_custom_variables) Set new width, print members.
387 (hash_value_alpha) Get width from var_freqs.
388 (compare_value_alpha_a) Ditto.
389 (compare_freq_alpha_a) Ditto.
390 (compare_freq_alpha_d) Ditto.
391 (dump_full) Get display format from var_freqs.
392 (dump_condensed) Ditto.
394 Mon Jun 19 22:07:13 2006 Ben Pfaff <blp@gnu.org>
396 * frequencies.q: (dump_full) Only put the first MAX_SHORT_STRING
397 bytes of string variables into the output cells, seeing as we only
399 (dump_condensed) Ditto.
401 Mon Jun 19 21:52:05 2006 Ben Pfaff <blp@gnu.org>
403 Fixes a bug reported by Greg Hunt <greg@firmansyah.com>.
405 * frequencies.q: (hsh_hash_bytes) We only copy the first
406 MAX_SHORT_STRING bytes of string variables, so we must only
407 compare that many bytes, even if the string variable is longer.
408 (compare_value_alpha_a) Ditto.
409 (compare_freq_alpha_a) Ditto.
410 (compare_freq_alpha_d) Ditto.
412 2006-05-11 Jason Stover <jhs@math.gcsu.edu>
414 * regression.q: Adjusted code to account for cache->coeff being a
415 pspp_linreg_coeff **.
417 Sun May 7 18:31:25 2006 Ben Pfaff <blp@gnu.org>
421 * aggregate.c (cmd_aggregate): Free default_dict before replacing
424 Sun May 7 17:09:19 2006 Ben Pfaff <blp@gnu.org>
426 * flip.c (flip_file): Check return value of pool_fclose().
428 Sat May 6 16:00:13 2006 Ben Pfaff <blp@gnu.org>
430 Get rid of `char *c' member in union value, for cleanliness.
432 * aggregate.c: (union agr_argument) New union.
433 (struct agr_var) Change element type of arg[] from union value to
435 (parse_aggregate_functions) Change local variable types likewise.
437 * autorecode.c: (union arc_value) New union.
438 (struct arc_item) Change "from" from union value to union
440 (recode) Change local variable from union value to union
442 (autorecode_trns_proc) Ditto.
443 (compare_alpha_value) Ditto.
444 (hash_alpha_value) Ditto.
445 (compare_numeric_value) Ditto.
446 (hash_numeric_value) Ditto.
447 (autorecode_proc_func) Ditto.
449 Sat May 6 10:43:33 2006 Ben Pfaff <blp@gnu.org>
451 Continue reforming procedure execution. In this phase, get rid of
452 the output code for SPLIT FILE groups in procedure.c, which really
453 shouldn't be doing any output. Move it into the individual
454 procedures instead. This also adds some flexibility.
456 * crosstabs.q (precalc): Call output_split_file_values().
458 * descriptives.c (calc_descriptives): Ditto.
460 * examine.q (run_examine): Ditto.
462 * frequencies.q (precalc): Ditto.
464 * oneway.q (run_oneway): Ditto.
466 * regression.q (run_regression): Ditto.
468 * t-test.q (calculate): Ditto.
470 Wed May 3 23:05:31 2006 Ben Pfaff <blp@gnu.org>
472 Continue reforming procedure execution. In this phase, get rid of
473 many global variables, consolidating procedure execution in
474 procedure.c. Encapsulate transformations in new "struct
475 trns_chain". Also, change implementation of N OF CASES, FILTER,
476 and PROCESS IF from special cases to transformations.
478 * aggregate.c (cmd_aggregate) Use discard_variables().
480 2006-04-28 Jason Stover <jhs@math.gcsu.edu>
482 * regression.q (regression_trns_resid_proc): Pass only the
483 variables used in the model to (*model->residual)().
485 * regression.q (regression_trns_pred_proc): Pass only the
486 variables used in the model to (*model->pred)().
488 2006-04-26 Jason Stover <jhs@math.gcsu.edu>
490 * regression.q: Added support for multiple transformations.
492 * regression.q (regression_trns_resid_proc): New function.
494 * regression.q (regression_trns_pred_proc): New function.
496 * regression.q (subcommand_save): Added support for saving
499 * regression.q (regression_trns_free): New function.
501 * regression.q (reg_get_name): New function.
503 * regression.q (reg_save_var): New function.
505 Tue Apr 25 13:18:56 2006 Ben Pfaff <blp@gnu.org>
507 * rank.q (parse_rank_function): Use SE instead of ME for parse
510 Tue Apr 25 13:16:28 2006 Ben Pfaff <blp@gnu.org>
512 * flip.c (flip_sink_write): Use snprintf() to simplify a bit of
515 2006-04-21 Jason Stover <jhs@math.gcsu.edu>
517 * regression.q (try_name): New function. (Partly copied from
518 try_name in descriptives.c.)
520 * regression.q (subcommand_save): Choose residual variable names
523 2006-04-20 Jason Stover <jhs@math.gcsu.edu>
525 * regression.q (cmd_regression): Moved call to subcommand_save()
526 outside multipass_procedure_with_splits().
528 * regression.q (regression_trns_proc): Fixed value counter n_vals
529 before calling *model->residual().
531 2006-04-19 Jason Stover <jhs@math.gcsu.edu>
533 * regression.q (regression_trns_proc): Fixed the look-up of the
536 2006-04-18 Jason Stover <jhs@math.gcsu.edu>
538 * regression.q (regression_trns_proc): Look up the residual
539 variable in the linear regression cache.
541 * regression.q (subcommand_save): Set the residual variable in the
542 linear regression cache.
544 2006-04-17 Jason Stover <jason@wonko.gcsu.edu>
546 * regression.q (regression_trns_proc): Accept case_idx as an int
547 to match the definition of trns_proc_func.
549 2006-04-17 Jason Stover <jhs@math.gcsu.edu>
551 * regression.q (regression_trns_proc): New function.
553 * regression.q (subcommand_save): Create variable residuals and
554 add a transformation to assign values to them. Also free the
555 linreg_cache if the SAVE command was not called. Removed the
556 casereading loop. Placed actual computation of residuals in
557 regression_trns_proc.
559 * regression.q (run_regression): Moved call to free
560 pspp_linreg_cache to subcommand_save.
562 Sat Apr 15 18:01:03 2006 Ben Pfaff <blp@gnu.org>
564 * examine.q (output_examine): Add casts to fix warnings.
566 2006-04-07 Jason Stover <jhs@math.gcsu.edu>
568 * regression.q (subcommand_save): New function.
570 2006-04-04 Jason Stover <jhs@math.gcsu.edu>
572 * regression.q: New function reg_has_categorical () to tell
573 whether a pspp_linreg_struct was made with any variables of type
576 * regression.q: (subcommand_export): Call
577 reg_print_categorical_encoding() only if the model uses any
578 categorical variables.
580 Mon Mar 27 16:00:42 2006 Ben Pfaff <blp@gnu.org>
582 * crosstabs.q: (output_pivot_table) Drop spurious space from
585 2006-03-15 Jason Stover <jhs@math.gcsu.edu>
587 * regression.q: Added custom syntax parser for VARIABLES subcommand
589 * regression.q: Moved most instructions for run_regression ()
590 inside the loop over dependent variables.
592 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
594 * Moved files from src directory