1 2008-05-29 John Darrington <john@darrington.wattle.id.au>
3 * examine.q: Fixed bug where incorrect levels of dereferencing
4 were applied to pointers.
6 2008-04-09 John Darrington <john@darrington.wattle.id.au>
8 * regression.q: Fix display of degrees of freedom.
10 2008-04-08 Jason Stover <jhs@math.gcsu.edu>
12 * regression.q (identify_indep_vars): Don't panic unless
15 2008-03-16 Ben Pfaff <blp@gnu.org>
17 Bug #22037. Thanks to John Darrington for reporting this bug.
19 * crosstabs.q (calc_general): Only the short string prefix of long
20 string variables are tabulated, so we must not copy or zero out
23 2008-03-10 Jason Stover <jhs@debs.hoobahooba.net>
25 * regression.q (run_regression): Removed code for EXPORT
26 subcommand. Remove use of coefficient 0 as the intercept.
28 2008-02-14 John Darrington <john@darrington.wattle.id.au>
30 * examine.q: Fixed counts of missing variables. Thanks to
31 Jason Stover for reporting this problem.
33 2008-01-02 John Darrington <john@darrington.wattle.id.au>
35 * binomial.c chisquare.c examine.q frequencies.q oneway.q regression.q : updated
36 all users of var_get_value_name to use replacement function var_append_value_name.
38 2007-12-07 Ben Pfaff <blp@gnu.org>
42 * crosstabs.q (precalc): Initialize data structures even if the
43 first case cannot be read.
45 * frequencies.q (precalc): Ditto.
47 2007-11-03 Ben Pfaff <blp@gnu.org>
49 Allow output files to overwrite input files (bug #21280).
51 * aggregate.c (cmd_aggregate): Manage file handle reference
54 * correlations.q (internal_cmd_frequencies): Ditto.
55 (cor_custom_matrix): Ditto.
57 * regression.q (regression_custom_export): Ditto.
58 (cmd_regression): Ditto.
60 2007-10-12 Ben Pfaff <blp@gnu.org>
62 * flip.c (flip_file): No need to conditionally substitute for
63 "fseeko" and "off_t" manually anymore, as gnulib takes care of it
66 2007-09-21 Jason Stover <jhs@wonko.gcsu.edu>
68 * regression.q (run_regression): Partial fix of memory leak, bug
71 2007-09-19 Ben Pfaff <blp@gnu.org>
75 * aggregate.c (cmd_aggregate): Destroy casereader consistently,
76 even if casereader fails.
78 * examine.q (run_examine): Ditto.
80 * glm.q (run_glm): Ditto.
82 * oneway.q (run_oneway): Ditto.
84 * regression.q (run_regression): Ditto.
86 * t-test.q (calculate): Ditto.
88 * descriptives.c (calc_descriptives): Ditto. Also avoid
89 gratuitous casereader_clone.
91 2007-09-13 Jason Stover <jhs@debs.hoobahooba.net>
93 * regression.q (cmd_regression): Move declaration of models in to
94 definition of cmd_regression.
96 * regression.q (run_regression): Free mom to fix memory leak.
98 2007-09-12 Ben Pfaff <blp@gnu.org>
100 * crosstabs.q (postcalc): Free sorted_tab and the structures that
101 it points to, to plug a memory leak. Fixes bug #20910. Thanks to
102 John Darrington for reporting this bug and for review.
104 2007-09-04 Ben Pfaff <blp@gnu.org>
106 * crosstabs.q (cmd_crosstabs): Free xtab and the structures that
107 it points to, to plug a memory leak. Fixes bug #18315.
109 2007-08-15 Jason Stover <jhs@wonko.gcsu.edu>
111 * regression.q (identify_indep_vars): Print an error if dependent
112 and independent variables are the same. Fixes bug 19819.
114 2007-08-12 Ben Pfaff <blp@gnu.org>
116 * flip.c: Drop use of dict_get_compacted_dict_index_to_case_index
117 and just use the ordinary case indexes. There seemed to be no
118 reason for the former method.
120 2007-08-03 Ben Pfaff <blp@gnu.org>
122 * rank.q (rank_cmd): Instead of sorting by SPLIT FILE vars, group
123 by them. Fixes bug #17239.
124 Reviewed by John Darrington.
126 2007-08-01 Ben Pfaff <blp@gnu.org>
128 Clean up handling of median, by treating it almost like any other
129 percentile. Fixes bug #17424. Thanks to John Darrington for
131 * frequencies.q (internal_cmd_frequencies): Fix handling of bit
132 masks for `stats' variable. If median is selected, turn it off
133 and add a 50th percentile.
134 (add_percentile): Simplify code a little.
135 (calc_stats): Drop special casing of median.
136 (dump_statistics): Ditto, except that we label the 50th percentile
137 as "50 (Median)" to make it clear that it's also the median.
139 2007-07-31 Ben Pfaff <blp@gnu.org>
141 Remove integer mode from FREQUENCIES and incidentally fix bug
142 #17421. Reviewed by John Darrington.
143 * frequencies.q (int_pool): Rename data_pool.
144 (gen_pool): Rename syntax_pool.
145 (enum FRQM_*): Removed.
146 (struct freq_tab): Removed `mode', `vector', `min', `max',
147 `out_of_range', `sysmis' members.
148 (calc): Delete support for integer mode.
150 (postprocess_freq_tab): Ditto.
151 (cleanup_freq_tab): Ditto.
152 (frq_custom_variables): Ditto.
154 2007-07-28 John Darrington <john@darrington.wattle.id.au>
156 * t-test.q: Moved the order in which groups are displayed in the
157 independent samples case, where a cut point is given.
159 2007-07-27 Ben Pfaff <blp@gnu.org>
161 * regression.q (run_regression): Move casereader_destroy call so
162 that it always gets called, not just if there was some valid
163 data. Fixes bug #19581.
164 Reviewed by Jason Stover.
166 2007-07-24 Ben Pfaff <blp@gnu.org>
168 * flip.c (struct flip_pgm): Remove `case_size' member (now
170 (cmd_flip): Pass var_cnt as number of cases instead of case_cnt,
171 to fix bug #20494. Don't assign to `case_size' member, which was
172 unused after assignment.
173 (build_dictionary): When NEWNAMES not used, get the number of
174 variables right, to fix bug #20493.
176 2007-07-10 Jason Stover <jhs@math.gcsu.edu>
178 * glm.q: Initial version of the GLM procedure.
180 2007-06-06 Ben Pfaff <blp@gnu.org>
182 Adapt case sources, sinks, and clients of procedure code to the
185 * aggregate.c: Simplify greatly since everything is more uniform
188 * autorecode.c: Adapt to new procedure code.
190 * chisquare.c: Ditto.
191 * crosstabs.q: Ditto.
192 * descriptives.c: Ditto.
194 * npar-summary.c: Ditto.
195 * frequencies.q: Ditto.
198 * regression.q: Ditto.
199 * sort-cases.c: Ditto.
202 * sort-criteria.c: Rewrite to output a struct case_ordering.
204 * flip.c: Rewrite to be a casereader.
206 * rank.q: Simplify greatly since casereaders are much more
207 flexible than what we had before.
209 2007-05-15 Jason Stover <jhs@math.gcsu.edu>
211 * regression.q (run_regression): Tell the user when the data
212 contain no valid cases.
214 2007-05-08 Jason Stover <jhs@math.gcsu.edu>
216 * regression.q: Partial fix of bug which caused a crash if
217 dependent variable and independent variable were the same.
219 2007-04-16 John Darrington <john@darrington.wattle.id.au>
221 * t-test.q: Changed the output width of reported counts and
222 degrees of freedom, to avoid truncating these values. Thanks
223 to Seth Woolley for reporting this problem. A proper fix involves
224 re-thinking the output driver.
226 2007-04-12 Jason Stover <jhs@math.gcsu.edu>
228 * regression.q (run_regression): Added if (n_data >0) to fix bug
231 2007-03-29 Jason Stover <jhs@math.gcsu.edu>
233 * regression.q (prepare_data): New function.
235 * regression.q (compute_moments): New function.
237 2007-03-18 Ben Pfaff <blp@gnu.org>
239 * crosstabs.q (static var write): Rename write_style to avoid
240 conflict with POSIX function of same name.
242 2007-03-16 Jason Stover <jhs@math.gcsu.edu>
244 * regression.q (run_regression): Added support for moments.
246 Sat Feb 17 08:16:00 2007 Ben Pfaff <blp@gnu.org>
248 * flip.c (flip_sink_create): Improve error message when temporary
249 file cannot be created.
251 Tue Feb 6 19:58:03 2007 Ben Pfaff <blp@gnu.org>
253 * flip.c (flip_file): Give better error message on end-of-file.
255 2007-02-04 Jason Stover <jhs@math.gcsu.edu>
257 * regression.q: Fixed p-value computation in the test for
258 individual regression coefficients.
260 Mon Jan 15 11:03:20 2007 Ben Pfaff <blp@gnu.org>
262 Fix bugs found by valgrind when --enable-debug is used with the
263 new case code. These bugs are hidden when the data set is small
264 enough to find in memory; if a bigger data set that would overflow
265 to disk were used, then data corruption would occur.
267 * chisquare.c (create_freq_hash): Pass free_freq_mutable_hash to
268 hsh_create as free function. Make copy of data put into hash.
270 * oneway.q (free_value): New function.
271 (run_oneway): Use free_value as arg to hsh_create. Make copy of
274 * rank.q (rank_cases): Don't access data in case after we've given
277 Tue Jan 9 19:16:11 2007 Ben Pfaff <blp@gnu.org>
281 * aggregate.c (parse_aggregate_functions) Initialize
284 Fri Dec 22 14:04:09 2006 Ben Pfaff <blp@gnu.org>
286 Simplify missing value handling.
288 * aggregate.c (struct agr_var): Remove `bool include_missing', add
289 `enum mv_class exclude'. Remove `int missing', add `bool
290 saw_missing'. Update users.
292 * descriptives.c (struct dsc_trns): Removed `int
293 include_user_missing', add `enum mv_class exclude'. Update users.
294 (struct dsc_proc): Ditto.
296 * examine.q: (static var value_is_missing): Rename
297 `exclude_values', change type to `enum mv_class'. Update users.
301 Fri Dec 22 19:22:18 WST 2006 John Darrington <john@darrington.wattle.id.au>
303 * frequencies.q : Fixed bug #17420, where the table bounds were overun
304 when /FORMAT=nolabels was given.
306 Wed Dec 20 18:45:31 WST 2006 John Darrington <john@darrington.wattle.id.au>
308 * binomial.c binomial.h : New files. Thanks to Jason Stover
309 for assistance with these.
311 * chisquare.c chisquare.h freq.c freq.h npar-summary.c npar-summary.h
312 npar.h npar.q: New files. Implementing NPAR TESTS.
314 * frequencies.q : Moved structure definitions into freq.[ch]
316 Sat Dec 16 22:26:44 2006 Ben Pfaff <blp@gnu.org>
318 Make it possible to pull cases from the active file with a
319 function call, instead of requiring indirection through a callback
322 * aggregate.c (cmd_aggregate): Take advantage of new procedure
324 (agr_to_active_file): Removed.
325 (presorted_agr_to_sysfile): Removed.
327 * autorecode.c (cmd_autorecode): Take advantage of new procedure
329 (autorecode_proc_func): Removed.
331 * flip.c (struct flip_pgm): New members to allow conformance with
332 new case_source_class interface.
333 (cmd_flip): Adapt to new case_source_class interface.
334 (flip_source_read): Ditto.
335 (flip_source_destroy): Ditto.
337 Sat Dec 16 12:54:27 2006 Ben Pfaff <blp@gnu.org>
339 * rank.q (rank_custom_variables): Allow grouping variables to be
340 strings. Fixes bug #18533. Thanks to John Darrington for review.
342 Sat Dec 9 18:47:51 2006 Ben Pfaff <blp@gnu.org>
344 * regression.q (is_depvar): Compare variable pointers instead of
347 Thu Dec 7 15:26:25 WST 2006 John Darrington <john@darrington.wattle.id.au>
349 * examine.q: Allocated the categorical values for the dependent and
350 independent variables, on the heap. Hence they can be of any width.
352 Wed Dec 6 21:14:26 2006 Ben Pfaff <blp@gnu.org>
354 * regression.q (reg_inserted): Compare variable pointers instead
357 Mon Dec 4 22:33:46 2006 Ben Pfaff <blp@gnu.org>
359 * crosstabs.q (insert_summary): Use var_to_string for labeling.
360 (output_pivot_table) Ditto.
363 * frequencies.q (setup_z_trns): Ditto.
365 (dump_condensed) Ditto.
366 (dump_statistics) Ditto.
369 Sun Nov 5 08:31:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
371 * t-test.q, oneway.q: Changed to use the new casefilter structure.
373 Sat Oct 14 16:52:28 2006 Ben Pfaff <blp@gnu.org>
375 * rank.q: (rank_sorted_casefile) Add some missing case_destroy()
376 calls to fix a memory leak.
378 Sun Oct 8 09:45:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
380 * rank.q: Plugged a small memory leak which occurred under error
383 Sat Oct 7 11:06:01 WST 2006 John Darrington <john@darrington.wattle.id.au>
385 * rank.q: Implemented most of the RANK command.
387 2006-07-14 Jason Stover <jhs@math.gcsu.edu>
389 * regression.q (run_regression): New function to move knowledge of
390 pspp_linreg_cache out of math/coefficient.[ch].
392 Sat Jul 1 17:41:46 2006 Ben Pfaff <blp@gnu.org>
394 Fix bug #11612, "q2c documentation does not agree with code".
396 * examine.q: Audit use of q2c "+" prefixes that indicate that a
397 command may appear multiple times.
399 * frequencies.q: Ditto.
403 * regression.q: Ditto.
407 Fri Jun 23 14:18:22 2006 Ben Pfaff <blp@gnu.org>
409 Support long string variables on FREQUENCIES, as
410 an extension when in enhanced algorithms mode. For Greg Hunt
411 <greg@firmansyah.com>.
413 * frequencies.q: (struct freq) Change `v' member from union value
414 to union value *. Update all references.
415 (struct var_freqs) Add width, print members to represent effective
416 variable width and display format.
417 (calc) Copy entire long string value into the hash table.
418 (frq_custom_variables) Set new width, print members.
419 (hash_value_alpha) Get width from var_freqs.
420 (compare_value_alpha_a) Ditto.
421 (compare_freq_alpha_a) Ditto.
422 (compare_freq_alpha_d) Ditto.
423 (dump_full) Get display format from var_freqs.
424 (dump_condensed) Ditto.
426 Mon Jun 19 22:07:13 2006 Ben Pfaff <blp@gnu.org>
428 * frequencies.q: (dump_full) Only put the first MAX_SHORT_STRING
429 bytes of string variables into the output cells, seeing as we only
431 (dump_condensed) Ditto.
433 Mon Jun 19 21:52:05 2006 Ben Pfaff <blp@gnu.org>
435 Fixes a bug reported by Greg Hunt <greg@firmansyah.com>.
437 * frequencies.q: (hsh_hash_bytes) We only copy the first
438 MAX_SHORT_STRING bytes of string variables, so we must only
439 compare that many bytes, even if the string variable is longer.
440 (compare_value_alpha_a) Ditto.
441 (compare_freq_alpha_a) Ditto.
442 (compare_freq_alpha_d) Ditto.
444 2006-05-11 Jason Stover <jhs@math.gcsu.edu>
446 * regression.q: Adjusted code to account for cache->coeff being a
447 pspp_linreg_coeff **.
449 Sun May 7 18:31:25 2006 Ben Pfaff <blp@gnu.org>
453 * aggregate.c (cmd_aggregate): Free default_dict before replacing
456 Sun May 7 17:09:19 2006 Ben Pfaff <blp@gnu.org>
458 * flip.c (flip_file): Check return value of pool_fclose().
460 Sat May 6 16:00:13 2006 Ben Pfaff <blp@gnu.org>
462 Get rid of `char *c' member in union value, for cleanliness.
464 * aggregate.c: (union agr_argument) New union.
465 (struct agr_var) Change element type of arg[] from union value to
467 (parse_aggregate_functions) Change local variable types likewise.
469 * autorecode.c: (union arc_value) New union.
470 (struct arc_item) Change "from" from union value to union
472 (recode) Change local variable from union value to union
474 (autorecode_trns_proc) Ditto.
475 (compare_alpha_value) Ditto.
476 (hash_alpha_value) Ditto.
477 (compare_numeric_value) Ditto.
478 (hash_numeric_value) Ditto.
479 (autorecode_proc_func) Ditto.
481 Sat May 6 10:43:33 2006 Ben Pfaff <blp@gnu.org>
483 Continue reforming procedure execution. In this phase, get rid of
484 the output code for SPLIT FILE groups in procedure.c, which really
485 shouldn't be doing any output. Move it into the individual
486 procedures instead. This also adds some flexibility.
488 * crosstabs.q (precalc): Call output_split_file_values().
490 * descriptives.c (calc_descriptives): Ditto.
492 * examine.q (run_examine): Ditto.
494 * frequencies.q (precalc): Ditto.
496 * oneway.q (run_oneway): Ditto.
498 * regression.q (run_regression): Ditto.
500 * t-test.q (calculate): Ditto.
502 Wed May 3 23:05:31 2006 Ben Pfaff <blp@gnu.org>
504 Continue reforming procedure execution. In this phase, get rid of
505 many global variables, consolidating procedure execution in
506 procedure.c. Encapsulate transformations in new "struct
507 trns_chain". Also, change implementation of N OF CASES, FILTER,
508 and PROCESS IF from special cases to transformations.
510 * aggregate.c (cmd_aggregate) Use discard_variables().
512 2006-04-28 Jason Stover <jhs@math.gcsu.edu>
514 * regression.q (regression_trns_resid_proc): Pass only the
515 variables used in the model to (*model->residual)().
517 * regression.q (regression_trns_pred_proc): Pass only the
518 variables used in the model to (*model->pred)().
520 2006-04-26 Jason Stover <jhs@math.gcsu.edu>
522 * regression.q: Added support for multiple transformations.
524 * regression.q (regression_trns_resid_proc): New function.
526 * regression.q (regression_trns_pred_proc): New function.
528 * regression.q (subcommand_save): Added support for saving
531 * regression.q (regression_trns_free): New function.
533 * regression.q (reg_get_name): New function.
535 * regression.q (reg_save_var): New function.
537 Tue Apr 25 13:18:56 2006 Ben Pfaff <blp@gnu.org>
539 * rank.q (parse_rank_function): Use SE instead of ME for parse
542 Tue Apr 25 13:16:28 2006 Ben Pfaff <blp@gnu.org>
544 * flip.c (flip_sink_write): Use snprintf() to simplify a bit of
547 2006-04-21 Jason Stover <jhs@math.gcsu.edu>
549 * regression.q (try_name): New function. (Partly copied from
550 try_name in descriptives.c.)
552 * regression.q (subcommand_save): Choose residual variable names
555 2006-04-20 Jason Stover <jhs@math.gcsu.edu>
557 * regression.q (cmd_regression): Moved call to subcommand_save()
558 outside multipass_procedure_with_splits().
560 * regression.q (regression_trns_proc): Fixed value counter n_vals
561 before calling *model->residual().
563 2006-04-19 Jason Stover <jhs@math.gcsu.edu>
565 * regression.q (regression_trns_proc): Fixed the look-up of the
568 2006-04-18 Jason Stover <jhs@math.gcsu.edu>
570 * regression.q (regression_trns_proc): Look up the residual
571 variable in the linear regression cache.
573 * regression.q (subcommand_save): Set the residual variable in the
574 linear regression cache.
576 2006-04-17 Jason Stover <jason@wonko.gcsu.edu>
578 * regression.q (regression_trns_proc): Accept case_idx as an int
579 to match the definition of trns_proc_func.
581 2006-04-17 Jason Stover <jhs@math.gcsu.edu>
583 * regression.q (regression_trns_proc): New function.
585 * regression.q (subcommand_save): Create variable residuals and
586 add a transformation to assign values to them. Also free the
587 linreg_cache if the SAVE command was not called. Removed the
588 casereading loop. Placed actual computation of residuals in
589 regression_trns_proc.
591 * regression.q (run_regression): Moved call to free
592 pspp_linreg_cache to subcommand_save.
594 Sat Apr 15 18:01:03 2006 Ben Pfaff <blp@gnu.org>
596 * examine.q (output_examine): Add casts to fix warnings.
598 2006-04-07 Jason Stover <jhs@math.gcsu.edu>
600 * regression.q (subcommand_save): New function.
602 2006-04-04 Jason Stover <jhs@math.gcsu.edu>
604 * regression.q: New function reg_has_categorical () to tell
605 whether a pspp_linreg_struct was made with any variables of type
608 * regression.q: (subcommand_export): Call
609 reg_print_categorical_encoding() only if the model uses any
610 categorical variables.
612 Mon Mar 27 16:00:42 2006 Ben Pfaff <blp@gnu.org>
614 * crosstabs.q: (output_pivot_table) Drop spurious space from
617 2006-03-15 Jason Stover <jhs@math.gcsu.edu>
619 * regression.q: Added custom syntax parser for VARIABLES subcommand
621 * regression.q: Moved most instructions for run_regression ()
622 inside the loop over dependent variables.
624 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
626 * Moved files from src directory