1 2008-07-22 Jason H Stover <jhs@math.gcsu.edu>
3 * glm.q (run_glm): Re-written to form covariance matrix rather
4 than store entire data set in memory.
6 * glm.q (data_pass_one): Renamed prepare_categories() to
7 data_pass_one(). Accumulate mean and variance.
9 2008-06-21 Jason Stover <jhs@math.gcsu.edu>
11 * regression.q (reg_stats_coeff): Use new accessor function
12 pspp_coeff_get_sd. Fixed bug 23567. No longer call compute_moments
13 in run_regression. Pass entire design_matrix to pspp_linreg ().
15 2008-05-29 John Darrington <john@darrington.wattle.id.au>
17 * examine.q: Fixed bug where incorrect levels of dereferencing
18 were applied to pointers.
20 2008-04-09 John Darrington <john@darrington.wattle.id.au>
22 * regression.q: Fix display of degrees of freedom.
24 2008-04-08 Jason Stover <jhs@math.gcsu.edu>
26 * regression.q (identify_indep_vars): Don't panic unless
29 2008-03-16 Ben Pfaff <blp@gnu.org>
31 Bug #22037. Thanks to John Darrington for reporting this bug.
33 * crosstabs.q (calc_general): Only the short string prefix of long
34 string variables are tabulated, so we must not copy or zero out
37 2008-03-10 Jason Stover <jhs@math.gcsu.edu>
39 * regression.q (run_regression): Removed code for EXPORT
40 subcommand. Remove use of coefficient 0 as the intercept.
42 2008-02-14 John Darrington <john@darrington.wattle.id.au>
44 * examine.q: Fixed counts of missing variables. Thanks to
45 Jason Stover for reporting this problem.
47 2008-01-02 John Darrington <john@darrington.wattle.id.au>
49 * binomial.c chisquare.c examine.q frequencies.q oneway.q regression.q : updated
50 all users of var_get_value_name to use replacement function var_append_value_name.
52 2007-12-07 Ben Pfaff <blp@gnu.org>
56 * crosstabs.q (precalc): Initialize data structures even if the
57 first case cannot be read.
59 * frequencies.q (precalc): Ditto.
61 2007-11-03 Ben Pfaff <blp@gnu.org>
63 Allow output files to overwrite input files (bug #21280).
65 * aggregate.c (cmd_aggregate): Manage file handle reference
68 * correlations.q (internal_cmd_frequencies): Ditto.
69 (cor_custom_matrix): Ditto.
71 * regression.q (regression_custom_export): Ditto.
72 (cmd_regression): Ditto.
74 2007-10-12 Ben Pfaff <blp@gnu.org>
76 * flip.c (flip_file): No need to conditionally substitute for
77 "fseeko" and "off_t" manually anymore, as gnulib takes care of it
80 2007-09-21 Jason Stover <jhs@wonko.gcsu.edu>
82 * regression.q (run_regression): Partial fix of memory leak, bug
85 2007-09-19 Ben Pfaff <blp@gnu.org>
89 * aggregate.c (cmd_aggregate): Destroy casereader consistently,
90 even if casereader fails.
92 * examine.q (run_examine): Ditto.
94 * glm.q (run_glm): Ditto.
96 * oneway.q (run_oneway): Ditto.
98 * regression.q (run_regression): Ditto.
100 * t-test.q (calculate): Ditto.
102 * descriptives.c (calc_descriptives): Ditto. Also avoid
103 gratuitous casereader_clone.
105 2007-09-13 Jason Stover <jhs@math.gcsu.edu>
107 * regression.q (cmd_regression): Move declaration of models in to
108 definition of cmd_regression.
110 * regression.q (run_regression): Free mom to fix memory leak.
112 2007-09-12 Ben Pfaff <blp@gnu.org>
114 * crosstabs.q (postcalc): Free sorted_tab and the structures that
115 it points to, to plug a memory leak. Fixes bug #20910. Thanks to
116 John Darrington for reporting this bug and for review.
118 2007-09-04 Ben Pfaff <blp@gnu.org>
120 * crosstabs.q (cmd_crosstabs): Free xtab and the structures that
121 it points to, to plug a memory leak. Fixes bug #18315.
123 2007-08-15 Jason Stover <jhs@wonko.gcsu.edu>
125 * regression.q (identify_indep_vars): Print an error if dependent
126 and independent variables are the same. Fixes bug 19819.
128 2007-08-12 Ben Pfaff <blp@gnu.org>
130 * flip.c: Drop use of dict_get_compacted_dict_index_to_case_index
131 and just use the ordinary case indexes. There seemed to be no
132 reason for the former method.
134 2007-08-03 Ben Pfaff <blp@gnu.org>
136 * rank.q (rank_cmd): Instead of sorting by SPLIT FILE vars, group
137 by them. Fixes bug #17239.
138 Reviewed by John Darrington.
140 2007-08-01 Ben Pfaff <blp@gnu.org>
142 Clean up handling of median, by treating it almost like any other
143 percentile. Fixes bug #17424. Thanks to John Darrington for
145 * frequencies.q (internal_cmd_frequencies): Fix handling of bit
146 masks for `stats' variable. If median is selected, turn it off
147 and add a 50th percentile.
148 (add_percentile): Simplify code a little.
149 (calc_stats): Drop special casing of median.
150 (dump_statistics): Ditto, except that we label the 50th percentile
151 as "50 (Median)" to make it clear that it's also the median.
153 2007-07-31 Ben Pfaff <blp@gnu.org>
155 Remove integer mode from FREQUENCIES and incidentally fix bug
156 #17421. Reviewed by John Darrington.
157 * frequencies.q (int_pool): Rename data_pool.
158 (gen_pool): Rename syntax_pool.
159 (enum FRQM_*): Removed.
160 (struct freq_tab): Removed `mode', `vector', `min', `max',
161 `out_of_range', `sysmis' members.
162 (calc): Delete support for integer mode.
164 (postprocess_freq_tab): Ditto.
165 (cleanup_freq_tab): Ditto.
166 (frq_custom_variables): Ditto.
168 2007-07-28 John Darrington <john@darrington.wattle.id.au>
170 * t-test.q: Moved the order in which groups are displayed in the
171 independent samples case, where a cut point is given.
173 2007-07-27 Ben Pfaff <blp@gnu.org>
175 * regression.q (run_regression): Move casereader_destroy call so
176 that it always gets called, not just if there was some valid
177 data. Fixes bug #19581.
178 Reviewed by Jason Stover.
180 2007-07-24 Ben Pfaff <blp@gnu.org>
182 * flip.c (struct flip_pgm): Remove `case_size' member (now
184 (cmd_flip): Pass var_cnt as number of cases instead of case_cnt,
185 to fix bug #20494. Don't assign to `case_size' member, which was
186 unused after assignment.
187 (build_dictionary): When NEWNAMES not used, get the number of
188 variables right, to fix bug #20493.
190 2007-07-10 Jason Stover <jhs@math.gcsu.edu>
192 * glm.q: Initial version of the GLM procedure.
194 2007-06-06 Ben Pfaff <blp@gnu.org>
196 Adapt case sources, sinks, and clients of procedure code to the
199 * aggregate.c: Simplify greatly since everything is more uniform
202 * autorecode.c: Adapt to new procedure code.
204 * chisquare.c: Ditto.
205 * crosstabs.q: Ditto.
206 * descriptives.c: Ditto.
208 * npar-summary.c: Ditto.
209 * frequencies.q: Ditto.
212 * regression.q: Ditto.
213 * sort-cases.c: Ditto.
216 * sort-criteria.c: Rewrite to output a struct case_ordering.
218 * flip.c: Rewrite to be a casereader.
220 * rank.q: Simplify greatly since casereaders are much more
221 flexible than what we had before.
223 2007-05-15 Jason Stover <jhs@math.gcsu.edu>
225 * regression.q (run_regression): Tell the user when the data
226 contain no valid cases.
228 2007-05-08 Jason Stover <jhs@math.gcsu.edu>
230 * regression.q: Partial fix of bug which caused a crash if
231 dependent variable and independent variable were the same.
233 2007-04-16 John Darrington <john@darrington.wattle.id.au>
235 * t-test.q: Changed the output width of reported counts and
236 degrees of freedom, to avoid truncating these values. Thanks
237 to Seth Woolley for reporting this problem. A proper fix involves
238 re-thinking the output driver.
240 2007-04-12 Jason Stover <jhs@math.gcsu.edu>
242 * regression.q (run_regression): Added if (n_data >0) to fix bug
245 2007-03-29 Jason Stover <jhs@math.gcsu.edu>
247 * regression.q (prepare_data): New function.
249 * regression.q (compute_moments): New function.
251 2007-03-18 Ben Pfaff <blp@gnu.org>
253 * crosstabs.q (static var write): Rename write_style to avoid
254 conflict with POSIX function of same name.
256 2007-03-16 Jason Stover <jhs@math.gcsu.edu>
258 * regression.q (run_regression): Added support for moments.
260 Sat Feb 17 08:16:00 2007 Ben Pfaff <blp@gnu.org>
262 * flip.c (flip_sink_create): Improve error message when temporary
263 file cannot be created.
265 Tue Feb 6 19:58:03 2007 Ben Pfaff <blp@gnu.org>
267 * flip.c (flip_file): Give better error message on end-of-file.
269 2007-02-04 Jason Stover <jhs@math.gcsu.edu>
271 * regression.q: Fixed p-value computation in the test for
272 individual regression coefficients.
274 Mon Jan 15 11:03:20 2007 Ben Pfaff <blp@gnu.org>
276 Fix bugs found by valgrind when --enable-debug is used with the
277 new case code. These bugs are hidden when the data set is small
278 enough to find in memory; if a bigger data set that would overflow
279 to disk were used, then data corruption would occur.
281 * chisquare.c (create_freq_hash): Pass free_freq_mutable_hash to
282 hsh_create as free function. Make copy of data put into hash.
284 * oneway.q (free_value): New function.
285 (run_oneway): Use free_value as arg to hsh_create. Make copy of
288 * rank.q (rank_cases): Don't access data in case after we've given
291 Tue Jan 9 19:16:11 2007 Ben Pfaff <blp@gnu.org>
295 * aggregate.c (parse_aggregate_functions) Initialize
298 Fri Dec 22 14:04:09 2006 Ben Pfaff <blp@gnu.org>
300 Simplify missing value handling.
302 * aggregate.c (struct agr_var): Remove `bool include_missing', add
303 `enum mv_class exclude'. Remove `int missing', add `bool
304 saw_missing'. Update users.
306 * descriptives.c (struct dsc_trns): Removed `int
307 include_user_missing', add `enum mv_class exclude'. Update users.
308 (struct dsc_proc): Ditto.
310 * examine.q: (static var value_is_missing): Rename
311 `exclude_values', change type to `enum mv_class'. Update users.
315 Fri Dec 22 19:22:18 WST 2006 John Darrington <john@darrington.wattle.id.au>
317 * frequencies.q : Fixed bug #17420, where the table bounds were overun
318 when /FORMAT=nolabels was given.
320 Wed Dec 20 18:45:31 WST 2006 John Darrington <john@darrington.wattle.id.au>
322 * binomial.c binomial.h : New files. Thanks to Jason Stover
323 for assistance with these.
325 * chisquare.c chisquare.h freq.c freq.h npar-summary.c npar-summary.h
326 npar.h npar.q: New files. Implementing NPAR TESTS.
328 * frequencies.q : Moved structure definitions into freq.[ch]
330 Sat Dec 16 22:26:44 2006 Ben Pfaff <blp@gnu.org>
332 Make it possible to pull cases from the active file with a
333 function call, instead of requiring indirection through a callback
336 * aggregate.c (cmd_aggregate): Take advantage of new procedure
338 (agr_to_active_file): Removed.
339 (presorted_agr_to_sysfile): Removed.
341 * autorecode.c (cmd_autorecode): Take advantage of new procedure
343 (autorecode_proc_func): Removed.
345 * flip.c (struct flip_pgm): New members to allow conformance with
346 new case_source_class interface.
347 (cmd_flip): Adapt to new case_source_class interface.
348 (flip_source_read): Ditto.
349 (flip_source_destroy): Ditto.
351 Sat Dec 16 12:54:27 2006 Ben Pfaff <blp@gnu.org>
353 * rank.q (rank_custom_variables): Allow grouping variables to be
354 strings. Fixes bug #18533. Thanks to John Darrington for review.
356 Sat Dec 9 18:47:51 2006 Ben Pfaff <blp@gnu.org>
358 * regression.q (is_depvar): Compare variable pointers instead of
361 Thu Dec 7 15:26:25 WST 2006 John Darrington <john@darrington.wattle.id.au>
363 * examine.q: Allocated the categorical values for the dependent and
364 independent variables, on the heap. Hence they can be of any width.
366 Wed Dec 6 21:14:26 2006 Ben Pfaff <blp@gnu.org>
368 * regression.q (reg_inserted): Compare variable pointers instead
371 Mon Dec 4 22:33:46 2006 Ben Pfaff <blp@gnu.org>
373 * crosstabs.q (insert_summary): Use var_to_string for labeling.
374 (output_pivot_table) Ditto.
377 * frequencies.q (setup_z_trns): Ditto.
379 (dump_condensed) Ditto.
380 (dump_statistics) Ditto.
383 Sun Nov 5 08:31:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
385 * t-test.q, oneway.q: Changed to use the new casefilter structure.
387 Sat Oct 14 16:52:28 2006 Ben Pfaff <blp@gnu.org>
389 * rank.q: (rank_sorted_casefile) Add some missing case_destroy()
390 calls to fix a memory leak.
392 Sun Oct 8 09:45:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
394 * rank.q: Plugged a small memory leak which occurred under error
397 Sat Oct 7 11:06:01 WST 2006 John Darrington <john@darrington.wattle.id.au>
399 * rank.q: Implemented most of the RANK command.
401 2006-07-14 Jason Stover <jhs@math.gcsu.edu>
403 * regression.q (run_regression): New function to move knowledge of
404 pspp_linreg_cache out of math/coefficient.[ch].
406 Sat Jul 1 17:41:46 2006 Ben Pfaff <blp@gnu.org>
408 Fix bug #11612, "q2c documentation does not agree with code".
410 * examine.q: Audit use of q2c "+" prefixes that indicate that a
411 command may appear multiple times.
413 * frequencies.q: Ditto.
417 * regression.q: Ditto.
421 Fri Jun 23 14:18:22 2006 Ben Pfaff <blp@gnu.org>
423 Support long string variables on FREQUENCIES, as
424 an extension when in enhanced algorithms mode. For Greg Hunt
425 <greg@firmansyah.com>.
427 * frequencies.q: (struct freq) Change `v' member from union value
428 to union value *. Update all references.
429 (struct var_freqs) Add width, print members to represent effective
430 variable width and display format.
431 (calc) Copy entire long string value into the hash table.
432 (frq_custom_variables) Set new width, print members.
433 (hash_value_alpha) Get width from var_freqs.
434 (compare_value_alpha_a) Ditto.
435 (compare_freq_alpha_a) Ditto.
436 (compare_freq_alpha_d) Ditto.
437 (dump_full) Get display format from var_freqs.
438 (dump_condensed) Ditto.
440 Mon Jun 19 22:07:13 2006 Ben Pfaff <blp@gnu.org>
442 * frequencies.q: (dump_full) Only put the first MAX_SHORT_STRING
443 bytes of string variables into the output cells, seeing as we only
445 (dump_condensed) Ditto.
447 Mon Jun 19 21:52:05 2006 Ben Pfaff <blp@gnu.org>
449 Fixes a bug reported by Greg Hunt <greg@firmansyah.com>.
451 * frequencies.q: (hsh_hash_bytes) We only copy the first
452 MAX_SHORT_STRING bytes of string variables, so we must only
453 compare that many bytes, even if the string variable is longer.
454 (compare_value_alpha_a) Ditto.
455 (compare_freq_alpha_a) Ditto.
456 (compare_freq_alpha_d) Ditto.
458 2006-05-11 Jason Stover <jhs@math.gcsu.edu>
460 * regression.q: Adjusted code to account for cache->coeff being a
461 pspp_linreg_coeff **.
463 Sun May 7 18:31:25 2006 Ben Pfaff <blp@gnu.org>
467 * aggregate.c (cmd_aggregate): Free default_dict before replacing
470 Sun May 7 17:09:19 2006 Ben Pfaff <blp@gnu.org>
472 * flip.c (flip_file): Check return value of pool_fclose().
474 Sat May 6 16:00:13 2006 Ben Pfaff <blp@gnu.org>
476 Get rid of `char *c' member in union value, for cleanliness.
478 * aggregate.c: (union agr_argument) New union.
479 (struct agr_var) Change element type of arg[] from union value to
481 (parse_aggregate_functions) Change local variable types likewise.
483 * autorecode.c: (union arc_value) New union.
484 (struct arc_item) Change "from" from union value to union
486 (recode) Change local variable from union value to union
488 (autorecode_trns_proc) Ditto.
489 (compare_alpha_value) Ditto.
490 (hash_alpha_value) Ditto.
491 (compare_numeric_value) Ditto.
492 (hash_numeric_value) Ditto.
493 (autorecode_proc_func) Ditto.
495 Sat May 6 10:43:33 2006 Ben Pfaff <blp@gnu.org>
497 Continue reforming procedure execution. In this phase, get rid of
498 the output code for SPLIT FILE groups in procedure.c, which really
499 shouldn't be doing any output. Move it into the individual
500 procedures instead. This also adds some flexibility.
502 * crosstabs.q (precalc): Call output_split_file_values().
504 * descriptives.c (calc_descriptives): Ditto.
506 * examine.q (run_examine): Ditto.
508 * frequencies.q (precalc): Ditto.
510 * oneway.q (run_oneway): Ditto.
512 * regression.q (run_regression): Ditto.
514 * t-test.q (calculate): Ditto.
516 Wed May 3 23:05:31 2006 Ben Pfaff <blp@gnu.org>
518 Continue reforming procedure execution. In this phase, get rid of
519 many global variables, consolidating procedure execution in
520 procedure.c. Encapsulate transformations in new "struct
521 trns_chain". Also, change implementation of N OF CASES, FILTER,
522 and PROCESS IF from special cases to transformations.
524 * aggregate.c (cmd_aggregate) Use discard_variables().
526 2006-04-28 Jason Stover <jhs@math.gcsu.edu>
528 * regression.q (regression_trns_resid_proc): Pass only the
529 variables used in the model to (*model->residual)().
531 * regression.q (regression_trns_pred_proc): Pass only the
532 variables used in the model to (*model->pred)().
534 2006-04-26 Jason Stover <jhs@math.gcsu.edu>
536 * regression.q: Added support for multiple transformations.
538 * regression.q (regression_trns_resid_proc): New function.
540 * regression.q (regression_trns_pred_proc): New function.
542 * regression.q (subcommand_save): Added support for saving
545 * regression.q (regression_trns_free): New function.
547 * regression.q (reg_get_name): New function.
549 * regression.q (reg_save_var): New function.
551 Tue Apr 25 13:18:56 2006 Ben Pfaff <blp@gnu.org>
553 * rank.q (parse_rank_function): Use SE instead of ME for parse
556 Tue Apr 25 13:16:28 2006 Ben Pfaff <blp@gnu.org>
558 * flip.c (flip_sink_write): Use snprintf() to simplify a bit of
561 2006-04-21 Jason Stover <jhs@math.gcsu.edu>
563 * regression.q (try_name): New function. (Partly copied from
564 try_name in descriptives.c.)
566 * regression.q (subcommand_save): Choose residual variable names
569 2006-04-20 Jason Stover <jhs@math.gcsu.edu>
571 * regression.q (cmd_regression): Moved call to subcommand_save()
572 outside multipass_procedure_with_splits().
574 * regression.q (regression_trns_proc): Fixed value counter n_vals
575 before calling *model->residual().
577 2006-04-19 Jason Stover <jhs@math.gcsu.edu>
579 * regression.q (regression_trns_proc): Fixed the look-up of the
582 2006-04-18 Jason Stover <jhs@math.gcsu.edu>
584 * regression.q (regression_trns_proc): Look up the residual
585 variable in the linear regression cache.
587 * regression.q (subcommand_save): Set the residual variable in the
588 linear regression cache.
590 2006-04-17 Jason Stover <jason@wonko.gcsu.edu>
592 * regression.q (regression_trns_proc): Accept case_idx as an int
593 to match the definition of trns_proc_func.
595 2006-04-17 Jason Stover <jhs@math.gcsu.edu>
597 * regression.q (regression_trns_proc): New function.
599 * regression.q (subcommand_save): Create variable residuals and
600 add a transformation to assign values to them. Also free the
601 linreg_cache if the SAVE command was not called. Removed the
602 casereading loop. Placed actual computation of residuals in
603 regression_trns_proc.
605 * regression.q (run_regression): Moved call to free
606 pspp_linreg_cache to subcommand_save.
608 Sat Apr 15 18:01:03 2006 Ben Pfaff <blp@gnu.org>
610 * examine.q (output_examine): Add casts to fix warnings.
612 2006-04-07 Jason Stover <jhs@math.gcsu.edu>
614 * regression.q (subcommand_save): New function.
616 2006-04-04 Jason Stover <jhs@math.gcsu.edu>
618 * regression.q: New function reg_has_categorical () to tell
619 whether a pspp_linreg_struct was made with any variables of type
622 * regression.q: (subcommand_export): Call
623 reg_print_categorical_encoding() only if the model uses any
624 categorical variables.
626 Mon Mar 27 16:00:42 2006 Ben Pfaff <blp@gnu.org>
628 * crosstabs.q: (output_pivot_table) Drop spurious space from
631 2006-03-15 Jason Stover <jhs@math.gcsu.edu>
633 * regression.q: Added custom syntax parser for VARIABLES subcommand
635 * regression.q: Moved most instructions for run_regression ()
636 inside the loop over dependent variables.
638 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
640 * Moved files from src directory