1 2008-07-24 Jason H Stover <jhs@math.gcsu.edu>
3 * glm.q (run_glm): Dropped weight argument.
5 2008-07-22 Jason H Stover <jhs@math.gcsu.edu>
7 * glm.q (run_glm): Re-written to form covariance matrix rather
8 than store entire data set in memory.
10 * glm.q (data_pass_one): Renamed prepare_categories() to
11 data_pass_one(). Accumulate mean and variance.
13 2008-06-21 Jason Stover <jhs@math.gcsu.edu>
15 * regression.q (reg_stats_coeff): Use new accessor function
16 pspp_coeff_get_sd. Fixed bug 23567. No longer call compute_moments
17 in run_regression. Pass entire design_matrix to pspp_linreg ().
19 2008-05-29 John Darrington <john@darrington.wattle.id.au>
21 * examine.q: Fixed bug where incorrect levels of dereferencing
22 were applied to pointers.
24 2008-04-09 John Darrington <john@darrington.wattle.id.au>
26 * regression.q: Fix display of degrees of freedom.
28 2008-04-08 Jason Stover <jhs@math.gcsu.edu>
30 * regression.q (identify_indep_vars): Don't panic unless
33 2008-03-16 Ben Pfaff <blp@gnu.org>
35 Bug #22037. Thanks to John Darrington for reporting this bug.
37 * crosstabs.q (calc_general): Only the short string prefix of long
38 string variables are tabulated, so we must not copy or zero out
41 2008-03-10 Jason Stover <jhs@math.gcsu.edu>
43 * regression.q (run_regression): Removed code for EXPORT
44 subcommand. Remove use of coefficient 0 as the intercept.
46 2008-02-14 John Darrington <john@darrington.wattle.id.au>
48 * examine.q: Fixed counts of missing variables. Thanks to
49 Jason Stover for reporting this problem.
51 2008-01-02 John Darrington <john@darrington.wattle.id.au>
53 * binomial.c chisquare.c examine.q frequencies.q oneway.q regression.q : updated
54 all users of var_get_value_name to use replacement function var_append_value_name.
56 2007-12-07 Ben Pfaff <blp@gnu.org>
60 * crosstabs.q (precalc): Initialize data structures even if the
61 first case cannot be read.
63 * frequencies.q (precalc): Ditto.
65 2007-11-03 Ben Pfaff <blp@gnu.org>
67 Allow output files to overwrite input files (bug #21280).
69 * aggregate.c (cmd_aggregate): Manage file handle reference
72 * correlations.q (internal_cmd_frequencies): Ditto.
73 (cor_custom_matrix): Ditto.
75 * regression.q (regression_custom_export): Ditto.
76 (cmd_regression): Ditto.
78 2007-10-12 Ben Pfaff <blp@gnu.org>
80 * flip.c (flip_file): No need to conditionally substitute for
81 "fseeko" and "off_t" manually anymore, as gnulib takes care of it
84 2007-09-21 Jason Stover <jhs@wonko.gcsu.edu>
86 * regression.q (run_regression): Partial fix of memory leak, bug
89 2007-09-19 Ben Pfaff <blp@gnu.org>
93 * aggregate.c (cmd_aggregate): Destroy casereader consistently,
94 even if casereader fails.
96 * examine.q (run_examine): Ditto.
98 * glm.q (run_glm): Ditto.
100 * oneway.q (run_oneway): Ditto.
102 * regression.q (run_regression): Ditto.
104 * t-test.q (calculate): Ditto.
106 * descriptives.c (calc_descriptives): Ditto. Also avoid
107 gratuitous casereader_clone.
109 2007-09-13 Jason Stover <jhs@math.gcsu.edu>
111 * regression.q (cmd_regression): Move declaration of models in to
112 definition of cmd_regression.
114 * regression.q (run_regression): Free mom to fix memory leak.
116 2007-09-12 Ben Pfaff <blp@gnu.org>
118 * crosstabs.q (postcalc): Free sorted_tab and the structures that
119 it points to, to plug a memory leak. Fixes bug #20910. Thanks to
120 John Darrington for reporting this bug and for review.
122 2007-09-04 Ben Pfaff <blp@gnu.org>
124 * crosstabs.q (cmd_crosstabs): Free xtab and the structures that
125 it points to, to plug a memory leak. Fixes bug #18315.
127 2007-08-15 Jason Stover <jhs@wonko.gcsu.edu>
129 * regression.q (identify_indep_vars): Print an error if dependent
130 and independent variables are the same. Fixes bug 19819.
132 2007-08-12 Ben Pfaff <blp@gnu.org>
134 * flip.c: Drop use of dict_get_compacted_dict_index_to_case_index
135 and just use the ordinary case indexes. There seemed to be no
136 reason for the former method.
138 2007-08-03 Ben Pfaff <blp@gnu.org>
140 * rank.q (rank_cmd): Instead of sorting by SPLIT FILE vars, group
141 by them. Fixes bug #17239.
142 Reviewed by John Darrington.
144 2007-08-01 Ben Pfaff <blp@gnu.org>
146 Clean up handling of median, by treating it almost like any other
147 percentile. Fixes bug #17424. Thanks to John Darrington for
149 * frequencies.q (internal_cmd_frequencies): Fix handling of bit
150 masks for `stats' variable. If median is selected, turn it off
151 and add a 50th percentile.
152 (add_percentile): Simplify code a little.
153 (calc_stats): Drop special casing of median.
154 (dump_statistics): Ditto, except that we label the 50th percentile
155 as "50 (Median)" to make it clear that it's also the median.
157 2007-07-31 Ben Pfaff <blp@gnu.org>
159 Remove integer mode from FREQUENCIES and incidentally fix bug
160 #17421. Reviewed by John Darrington.
161 * frequencies.q (int_pool): Rename data_pool.
162 (gen_pool): Rename syntax_pool.
163 (enum FRQM_*): Removed.
164 (struct freq_tab): Removed `mode', `vector', `min', `max',
165 `out_of_range', `sysmis' members.
166 (calc): Delete support for integer mode.
168 (postprocess_freq_tab): Ditto.
169 (cleanup_freq_tab): Ditto.
170 (frq_custom_variables): Ditto.
172 2007-07-28 John Darrington <john@darrington.wattle.id.au>
174 * t-test.q: Moved the order in which groups are displayed in the
175 independent samples case, where a cut point is given.
177 2007-07-27 Ben Pfaff <blp@gnu.org>
179 * regression.q (run_regression): Move casereader_destroy call so
180 that it always gets called, not just if there was some valid
181 data. Fixes bug #19581.
182 Reviewed by Jason Stover.
184 2007-07-24 Ben Pfaff <blp@gnu.org>
186 * flip.c (struct flip_pgm): Remove `case_size' member (now
188 (cmd_flip): Pass var_cnt as number of cases instead of case_cnt,
189 to fix bug #20494. Don't assign to `case_size' member, which was
190 unused after assignment.
191 (build_dictionary): When NEWNAMES not used, get the number of
192 variables right, to fix bug #20493.
194 2007-07-10 Jason Stover <jhs@math.gcsu.edu>
196 * glm.q: Initial version of the GLM procedure.
198 2007-06-06 Ben Pfaff <blp@gnu.org>
200 Adapt case sources, sinks, and clients of procedure code to the
203 * aggregate.c: Simplify greatly since everything is more uniform
206 * autorecode.c: Adapt to new procedure code.
208 * chisquare.c: Ditto.
209 * crosstabs.q: Ditto.
210 * descriptives.c: Ditto.
212 * npar-summary.c: Ditto.
213 * frequencies.q: Ditto.
216 * regression.q: Ditto.
217 * sort-cases.c: Ditto.
220 * sort-criteria.c: Rewrite to output a struct case_ordering.
222 * flip.c: Rewrite to be a casereader.
224 * rank.q: Simplify greatly since casereaders are much more
225 flexible than what we had before.
227 2007-05-15 Jason Stover <jhs@math.gcsu.edu>
229 * regression.q (run_regression): Tell the user when the data
230 contain no valid cases.
232 2007-05-08 Jason Stover <jhs@math.gcsu.edu>
234 * regression.q: Partial fix of bug which caused a crash if
235 dependent variable and independent variable were the same.
237 2007-04-16 John Darrington <john@darrington.wattle.id.au>
239 * t-test.q: Changed the output width of reported counts and
240 degrees of freedom, to avoid truncating these values. Thanks
241 to Seth Woolley for reporting this problem. A proper fix involves
242 re-thinking the output driver.
244 2007-04-12 Jason Stover <jhs@math.gcsu.edu>
246 * regression.q (run_regression): Added if (n_data >0) to fix bug
249 2007-03-29 Jason Stover <jhs@math.gcsu.edu>
251 * regression.q (prepare_data): New function.
253 * regression.q (compute_moments): New function.
255 2007-03-18 Ben Pfaff <blp@gnu.org>
257 * crosstabs.q (static var write): Rename write_style to avoid
258 conflict with POSIX function of same name.
260 2007-03-16 Jason Stover <jhs@math.gcsu.edu>
262 * regression.q (run_regression): Added support for moments.
264 Sat Feb 17 08:16:00 2007 Ben Pfaff <blp@gnu.org>
266 * flip.c (flip_sink_create): Improve error message when temporary
267 file cannot be created.
269 Tue Feb 6 19:58:03 2007 Ben Pfaff <blp@gnu.org>
271 * flip.c (flip_file): Give better error message on end-of-file.
273 2007-02-04 Jason Stover <jhs@math.gcsu.edu>
275 * regression.q: Fixed p-value computation in the test for
276 individual regression coefficients.
278 Mon Jan 15 11:03:20 2007 Ben Pfaff <blp@gnu.org>
280 Fix bugs found by valgrind when --enable-debug is used with the
281 new case code. These bugs are hidden when the data set is small
282 enough to find in memory; if a bigger data set that would overflow
283 to disk were used, then data corruption would occur.
285 * chisquare.c (create_freq_hash): Pass free_freq_mutable_hash to
286 hsh_create as free function. Make copy of data put into hash.
288 * oneway.q (free_value): New function.
289 (run_oneway): Use free_value as arg to hsh_create. Make copy of
292 * rank.q (rank_cases): Don't access data in case after we've given
295 Tue Jan 9 19:16:11 2007 Ben Pfaff <blp@gnu.org>
299 * aggregate.c (parse_aggregate_functions) Initialize
302 Fri Dec 22 14:04:09 2006 Ben Pfaff <blp@gnu.org>
304 Simplify missing value handling.
306 * aggregate.c (struct agr_var): Remove `bool include_missing', add
307 `enum mv_class exclude'. Remove `int missing', add `bool
308 saw_missing'. Update users.
310 * descriptives.c (struct dsc_trns): Removed `int
311 include_user_missing', add `enum mv_class exclude'. Update users.
312 (struct dsc_proc): Ditto.
314 * examine.q: (static var value_is_missing): Rename
315 `exclude_values', change type to `enum mv_class'. Update users.
319 Fri Dec 22 19:22:18 WST 2006 John Darrington <john@darrington.wattle.id.au>
321 * frequencies.q : Fixed bug #17420, where the table bounds were overun
322 when /FORMAT=nolabels was given.
324 Wed Dec 20 18:45:31 WST 2006 John Darrington <john@darrington.wattle.id.au>
326 * binomial.c binomial.h : New files. Thanks to Jason Stover
327 for assistance with these.
329 * chisquare.c chisquare.h freq.c freq.h npar-summary.c npar-summary.h
330 npar.h npar.q: New files. Implementing NPAR TESTS.
332 * frequencies.q : Moved structure definitions into freq.[ch]
334 Sat Dec 16 22:26:44 2006 Ben Pfaff <blp@gnu.org>
336 Make it possible to pull cases from the active file with a
337 function call, instead of requiring indirection through a callback
340 * aggregate.c (cmd_aggregate): Take advantage of new procedure
342 (agr_to_active_file): Removed.
343 (presorted_agr_to_sysfile): Removed.
345 * autorecode.c (cmd_autorecode): Take advantage of new procedure
347 (autorecode_proc_func): Removed.
349 * flip.c (struct flip_pgm): New members to allow conformance with
350 new case_source_class interface.
351 (cmd_flip): Adapt to new case_source_class interface.
352 (flip_source_read): Ditto.
353 (flip_source_destroy): Ditto.
355 Sat Dec 16 12:54:27 2006 Ben Pfaff <blp@gnu.org>
357 * rank.q (rank_custom_variables): Allow grouping variables to be
358 strings. Fixes bug #18533. Thanks to John Darrington for review.
360 Sat Dec 9 18:47:51 2006 Ben Pfaff <blp@gnu.org>
362 * regression.q (is_depvar): Compare variable pointers instead of
365 Thu Dec 7 15:26:25 WST 2006 John Darrington <john@darrington.wattle.id.au>
367 * examine.q: Allocated the categorical values for the dependent and
368 independent variables, on the heap. Hence they can be of any width.
370 Wed Dec 6 21:14:26 2006 Ben Pfaff <blp@gnu.org>
372 * regression.q (reg_inserted): Compare variable pointers instead
375 Mon Dec 4 22:33:46 2006 Ben Pfaff <blp@gnu.org>
377 * crosstabs.q (insert_summary): Use var_to_string for labeling.
378 (output_pivot_table) Ditto.
381 * frequencies.q (setup_z_trns): Ditto.
383 (dump_condensed) Ditto.
384 (dump_statistics) Ditto.
387 Sun Nov 5 08:31:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
389 * t-test.q, oneway.q: Changed to use the new casefilter structure.
391 Sat Oct 14 16:52:28 2006 Ben Pfaff <blp@gnu.org>
393 * rank.q: (rank_sorted_casefile) Add some missing case_destroy()
394 calls to fix a memory leak.
396 Sun Oct 8 09:45:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
398 * rank.q: Plugged a small memory leak which occurred under error
401 Sat Oct 7 11:06:01 WST 2006 John Darrington <john@darrington.wattle.id.au>
403 * rank.q: Implemented most of the RANK command.
405 2006-07-14 Jason Stover <jhs@math.gcsu.edu>
407 * regression.q (run_regression): New function to move knowledge of
408 pspp_linreg_cache out of math/coefficient.[ch].
410 Sat Jul 1 17:41:46 2006 Ben Pfaff <blp@gnu.org>
412 Fix bug #11612, "q2c documentation does not agree with code".
414 * examine.q: Audit use of q2c "+" prefixes that indicate that a
415 command may appear multiple times.
417 * frequencies.q: Ditto.
421 * regression.q: Ditto.
425 Fri Jun 23 14:18:22 2006 Ben Pfaff <blp@gnu.org>
427 Support long string variables on FREQUENCIES, as
428 an extension when in enhanced algorithms mode. For Greg Hunt
429 <greg@firmansyah.com>.
431 * frequencies.q: (struct freq) Change `v' member from union value
432 to union value *. Update all references.
433 (struct var_freqs) Add width, print members to represent effective
434 variable width and display format.
435 (calc) Copy entire long string value into the hash table.
436 (frq_custom_variables) Set new width, print members.
437 (hash_value_alpha) Get width from var_freqs.
438 (compare_value_alpha_a) Ditto.
439 (compare_freq_alpha_a) Ditto.
440 (compare_freq_alpha_d) Ditto.
441 (dump_full) Get display format from var_freqs.
442 (dump_condensed) Ditto.
444 Mon Jun 19 22:07:13 2006 Ben Pfaff <blp@gnu.org>
446 * frequencies.q: (dump_full) Only put the first MAX_SHORT_STRING
447 bytes of string variables into the output cells, seeing as we only
449 (dump_condensed) Ditto.
451 Mon Jun 19 21:52:05 2006 Ben Pfaff <blp@gnu.org>
453 Fixes a bug reported by Greg Hunt <greg@firmansyah.com>.
455 * frequencies.q: (hsh_hash_bytes) We only copy the first
456 MAX_SHORT_STRING bytes of string variables, so we must only
457 compare that many bytes, even if the string variable is longer.
458 (compare_value_alpha_a) Ditto.
459 (compare_freq_alpha_a) Ditto.
460 (compare_freq_alpha_d) Ditto.
462 2006-05-11 Jason Stover <jhs@math.gcsu.edu>
464 * regression.q: Adjusted code to account for cache->coeff being a
465 pspp_linreg_coeff **.
467 Sun May 7 18:31:25 2006 Ben Pfaff <blp@gnu.org>
471 * aggregate.c (cmd_aggregate): Free default_dict before replacing
474 Sun May 7 17:09:19 2006 Ben Pfaff <blp@gnu.org>
476 * flip.c (flip_file): Check return value of pool_fclose().
478 Sat May 6 16:00:13 2006 Ben Pfaff <blp@gnu.org>
480 Get rid of `char *c' member in union value, for cleanliness.
482 * aggregate.c: (union agr_argument) New union.
483 (struct agr_var) Change element type of arg[] from union value to
485 (parse_aggregate_functions) Change local variable types likewise.
487 * autorecode.c: (union arc_value) New union.
488 (struct arc_item) Change "from" from union value to union
490 (recode) Change local variable from union value to union
492 (autorecode_trns_proc) Ditto.
493 (compare_alpha_value) Ditto.
494 (hash_alpha_value) Ditto.
495 (compare_numeric_value) Ditto.
496 (hash_numeric_value) Ditto.
497 (autorecode_proc_func) Ditto.
499 Sat May 6 10:43:33 2006 Ben Pfaff <blp@gnu.org>
501 Continue reforming procedure execution. In this phase, get rid of
502 the output code for SPLIT FILE groups in procedure.c, which really
503 shouldn't be doing any output. Move it into the individual
504 procedures instead. This also adds some flexibility.
506 * crosstabs.q (precalc): Call output_split_file_values().
508 * descriptives.c (calc_descriptives): Ditto.
510 * examine.q (run_examine): Ditto.
512 * frequencies.q (precalc): Ditto.
514 * oneway.q (run_oneway): Ditto.
516 * regression.q (run_regression): Ditto.
518 * t-test.q (calculate): Ditto.
520 Wed May 3 23:05:31 2006 Ben Pfaff <blp@gnu.org>
522 Continue reforming procedure execution. In this phase, get rid of
523 many global variables, consolidating procedure execution in
524 procedure.c. Encapsulate transformations in new "struct
525 trns_chain". Also, change implementation of N OF CASES, FILTER,
526 and PROCESS IF from special cases to transformations.
528 * aggregate.c (cmd_aggregate) Use discard_variables().
530 2006-04-28 Jason Stover <jhs@math.gcsu.edu>
532 * regression.q (regression_trns_resid_proc): Pass only the
533 variables used in the model to (*model->residual)().
535 * regression.q (regression_trns_pred_proc): Pass only the
536 variables used in the model to (*model->pred)().
538 2006-04-26 Jason Stover <jhs@math.gcsu.edu>
540 * regression.q: Added support for multiple transformations.
542 * regression.q (regression_trns_resid_proc): New function.
544 * regression.q (regression_trns_pred_proc): New function.
546 * regression.q (subcommand_save): Added support for saving
549 * regression.q (regression_trns_free): New function.
551 * regression.q (reg_get_name): New function.
553 * regression.q (reg_save_var): New function.
555 Tue Apr 25 13:18:56 2006 Ben Pfaff <blp@gnu.org>
557 * rank.q (parse_rank_function): Use SE instead of ME for parse
560 Tue Apr 25 13:16:28 2006 Ben Pfaff <blp@gnu.org>
562 * flip.c (flip_sink_write): Use snprintf() to simplify a bit of
565 2006-04-21 Jason Stover <jhs@math.gcsu.edu>
567 * regression.q (try_name): New function. (Partly copied from
568 try_name in descriptives.c.)
570 * regression.q (subcommand_save): Choose residual variable names
573 2006-04-20 Jason Stover <jhs@math.gcsu.edu>
575 * regression.q (cmd_regression): Moved call to subcommand_save()
576 outside multipass_procedure_with_splits().
578 * regression.q (regression_trns_proc): Fixed value counter n_vals
579 before calling *model->residual().
581 2006-04-19 Jason Stover <jhs@math.gcsu.edu>
583 * regression.q (regression_trns_proc): Fixed the look-up of the
586 2006-04-18 Jason Stover <jhs@math.gcsu.edu>
588 * regression.q (regression_trns_proc): Look up the residual
589 variable in the linear regression cache.
591 * regression.q (subcommand_save): Set the residual variable in the
592 linear regression cache.
594 2006-04-17 Jason Stover <jason@wonko.gcsu.edu>
596 * regression.q (regression_trns_proc): Accept case_idx as an int
597 to match the definition of trns_proc_func.
599 2006-04-17 Jason Stover <jhs@math.gcsu.edu>
601 * regression.q (regression_trns_proc): New function.
603 * regression.q (subcommand_save): Create variable residuals and
604 add a transformation to assign values to them. Also free the
605 linreg_cache if the SAVE command was not called. Removed the
606 casereading loop. Placed actual computation of residuals in
607 regression_trns_proc.
609 * regression.q (run_regression): Moved call to free
610 pspp_linreg_cache to subcommand_save.
612 Sat Apr 15 18:01:03 2006 Ben Pfaff <blp@gnu.org>
614 * examine.q (output_examine): Add casts to fix warnings.
616 2006-04-07 Jason Stover <jhs@math.gcsu.edu>
618 * regression.q (subcommand_save): New function.
620 2006-04-04 Jason Stover <jhs@math.gcsu.edu>
622 * regression.q: New function reg_has_categorical () to tell
623 whether a pspp_linreg_struct was made with any variables of type
626 * regression.q: (subcommand_export): Call
627 reg_print_categorical_encoding() only if the model uses any
628 categorical variables.
630 Mon Mar 27 16:00:42 2006 Ben Pfaff <blp@gnu.org>
632 * crosstabs.q: (output_pivot_table) Drop spurious space from
635 2006-03-15 Jason Stover <jhs@math.gcsu.edu>
637 * regression.q: Added custom syntax parser for VARIABLES subcommand
639 * regression.q: Moved most instructions for run_regression ()
640 inside the loop over dependent variables.
642 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
644 * Moved files from src directory