X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Faggregate.c;h=0d181bd45e3999ab1c45b8267e713f6cf1aeb797;hb=5c3291dc396b795696e94f47780308fd7ace6fc4;hp=e1dcd12342ff2b42ff94dcbf472affeb58bb51c1;hpb=65be63912fedc9819f76099b1ed52b189fc5fc03;p=pspp-builds.git diff --git a/src/language/stats/aggregate.c b/src/language/stats/aggregate.c index e1dcd123..0d181bd4 100644 --- a/src/language/stats/aggregate.c +++ b/src/language/stats/aggregate.c @@ -1,5 +1,5 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 1997-9, 2000, 2006, 2008 Free Software Foundation, Inc. + Copyright (C) 1997-9, 2000, 2006, 2008, 2009 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -19,7 +19,6 @@ #include #include -#include #include #include #include @@ -29,6 +28,7 @@ #include #include #include +#include #include #include #include @@ -143,10 +143,10 @@ enum missing_treatment struct agr_proc { /* Break variables. */ - struct case_ordering *sort; /* Sort criteria (break variable). */ + struct subcase sort; /* Sort criteria (break variables). */ const struct variable **break_vars; /* Break variables. */ size_t break_var_cnt; /* Number of break variables. */ - struct ccase break_case; /* Last values of break variables. */ + struct ccase *break_case; /* Last values of break variables. */ enum missing_treatment missing; /* How to treat missing values. */ struct agr_var *agr_vars; /* First aggregate variable. */ @@ -187,10 +187,11 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) memset(&agr, 0 , sizeof (agr)); agr.missing = ITEMWISE; - case_nullify (&agr.break_case); + agr.break_case = NULL; agr.dict = dict_create (); agr.src_dict = dict; + subcase_init_empty (&agr.sort); dict_set_label (agr.dict, dict_get_label (dict)); dict_set_documents (agr.dict, dict_get_documents (dict)); @@ -229,13 +230,10 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) int i; lex_match (lexer, '='); - agr.sort = parse_case_ordering (lexer, dict, - - &saw_direction); - if (agr.sort == NULL) + if (!parse_sort_criteria (lexer, dict, &agr.sort, &agr.break_vars, + &saw_direction)) goto error; - case_ordering_get_vars (agr.sort, - &agr.break_vars, &agr.break_var_cnt); + agr.break_var_cnt = subcase_get_n_fields (&agr.sort); for (i = 0; i < agr.break_var_cnt; i++) dict_clone_var_assert (agr.dict, agr.break_vars[i], @@ -276,7 +274,7 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) so TEMPORARY is moot. */ proc_cancel_temporary_transformations (ds); proc_discard_output (ds); - output = autopaging_writer_create (dict_get_next_value_idx (agr.dict)); + output = autopaging_writer_create (dict_get_proto (agr.dict)); } else { @@ -286,10 +284,10 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) } input = proc_open (ds); - if (agr.sort != NULL && !presorted) + if (!subcase_is_empty (&agr.sort) && !presorted) { - input = sort_execute (input, agr.sort); - agr.sort = NULL; + input = sort_execute (input, &agr.sort); + subcase_clear (&agr.sort); } for (grouper = casegrouper_create_vars (input, agr.break_vars, @@ -297,18 +295,17 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) casegrouper_get_next_group (grouper, &group); casereader_destroy (group)) { - struct ccase c; - - if (!casereader_peek (group, 0, &c)) + struct ccase *c = casereader_peek (group, 0); + if (c == NULL) { casereader_destroy (group); continue; } - initialize_aggregate_info (&agr, &c); - case_destroy (&c); + initialize_aggregate_info (&agr, c); + case_unref (c); - for (; casereader_read (group, &c); case_destroy (&c)) - accumulate_aggregate_info (&agr, &c); + for (; (c = casereader_read (group)) != NULL; case_unref (c)) + accumulate_aggregate_info (&agr, c); dump_aggregate_info (&agr, output); } if (!casegrouper_destroy (grouper)) @@ -694,9 +691,9 @@ agr_destroy (struct agr_proc *agr) { struct agr_var *iter, *next; - case_ordering_destroy (agr->sort); + subcase_destroy (&agr->sort); free (agr->break_vars); - case_destroy (&agr->break_case); + case_unref (agr->break_case); for (iter = agr->agr_vars; iter; iter = next) { next = iter->next; @@ -772,20 +769,20 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) case MEDIAN: { double wv ; - struct ccase cout; - case_create (&cout, 2); + struct ccase *cout; + + cout = case_create (casewriter_get_proto (iter->writer)); - case_data_rw (&cout, iter->subject)->f = - case_data (input, iter->src)->f; + case_data_rw (cout, iter->subject)->f + = case_data (input, iter->src)->f; wv = dict_get_case_weight (agr->src_dict, input, NULL); - case_data_rw (&cout, iter->weight)->f = wv; + case_data_rw (cout, iter->weight)->f = wv; iter->cc += wv; - casewriter_write (iter->writer, &cout); - case_destroy (&cout); + casewriter_write (iter->writer, cout); } break; case SD: @@ -796,8 +793,8 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) iter->int1 = 1; break; case MAX | FSTRING: - if (memcmp (iter->string, v->s, src_width) < 0) - memcpy (iter->string, v->s, src_width); + if (memcmp (iter->string, value_str (v, src_width), src_width) < 0) + memcpy (iter->string, value_str (v, src_width), src_width); iter->int1 = 1; break; case MIN: @@ -805,8 +802,8 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) iter->int1 = 1; break; case MIN | FSTRING: - if (memcmp (iter->string, v->s, src_width) > 0) - memcpy (iter->string, v->s, src_width); + if (memcmp (iter->string, value_str (v, src_width), src_width) > 0) + memcpy (iter->string, value_str (v, src_width), src_width); iter->int1 = 1; break; case FGT: @@ -817,7 +814,8 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) break; case FGT | FSTRING: case PGT | FSTRING: - if (memcmp (iter->arg[0].c, v->s, src_width) < 0) + if (memcmp (iter->arg[0].c, + value_str (v, src_width), src_width) < 0) iter->dbl[0] += weight; iter->dbl[1] += weight; break; @@ -829,7 +827,8 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) break; case FLT | FSTRING: case PLT | FSTRING: - if (memcmp (iter->arg[0].c, v->s, src_width) > 0) + if (memcmp (iter->arg[0].c, + value_str (v, src_width), src_width) > 0) iter->dbl[0] += weight; iter->dbl[1] += weight; break; @@ -841,8 +840,10 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) break; case FIN | FSTRING: case PIN | FSTRING: - if (memcmp (iter->arg[0].c, v->s, src_width) <= 0 - && memcmp (iter->arg[1].c, v->s, src_width) >= 0) + if (memcmp (iter->arg[0].c, + value_str (v, src_width), src_width) <= 0 + && memcmp (iter->arg[1].c, + value_str (v, src_width), src_width) >= 0) iter->dbl[0] += weight; iter->dbl[1] += weight; break; @@ -854,8 +855,10 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) break; case FOUT | FSTRING: case POUT | FSTRING: - if (memcmp (iter->arg[0].c, v->s, src_width) > 0 - || memcmp (iter->arg[1].c, v->s, src_width) < 0) + if (memcmp (iter->arg[0].c, + value_str (v, src_width), src_width) > 0 + || memcmp (iter->arg[1].c, + value_str (v, src_width), src_width) < 0) iter->dbl[0] += weight; iter->dbl[1] += weight; break; @@ -877,7 +880,7 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) case FIRST | FSTRING: if (iter->int1 == 0) { - memcpy (iter->string, v->s, src_width); + memcpy (iter->string, value_str (v, src_width), src_width); iter->int1 = 1; } break; @@ -886,7 +889,7 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) iter->int1 = 1; break; case LAST | FSTRING: - memcpy (iter->string, v->s, src_width); + memcpy (iter->string, value_str (v, src_width), src_width); iter->int1 = 1; break; case NMISS: @@ -918,9 +921,7 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) static void dump_aggregate_info (struct agr_proc *agr, struct casewriter *output) { - struct ccase c; - - case_create (&c, dict_get_next_value_idx (agr->dict)); + struct ccase *c = case_create (dict_get_proto (agr->dict)); { int value_idx = 0; @@ -929,11 +930,10 @@ dump_aggregate_info (struct agr_proc *agr, struct casewriter *output) for (i = 0; i < agr->break_var_cnt; i++) { const struct variable *v = agr->break_vars[i]; - size_t value_cnt = var_get_value_cnt (v); - memcpy (case_data_rw_idx (&c, value_idx), - case_data (&agr->break_case, v), - sizeof (union value) * value_cnt); - value_idx += value_cnt; + value_copy (case_data_rw_idx (c, value_idx), + case_data (agr->break_case, v), + var_get_width (v)); + value_idx++; } } @@ -942,20 +942,15 @@ dump_aggregate_info (struct agr_proc *agr, struct casewriter *output) for (i = agr->agr_vars; i; i = i->next) { - union value *v = case_data_rw (&c, i->dest); - + union value *v = case_data_rw (c, i->dest); + int width = var_get_width (i->dest); if (agr->missing == COLUMNWISE && i->saw_missing && (i->function & FUNC) != N && (i->function & FUNC) != NU && (i->function & FUNC) != NMISS && (i->function & FUNC) != NUMISS) { - if (var_is_alpha (i->dest)) - memset (v->s, ' ', var_get_width (i->dest)); - else - v->f = SYSMIS; - + value_set_missing (v, width); casewriter_destroy (i->writer); - continue; } @@ -1006,9 +1001,9 @@ dump_aggregate_info (struct agr_proc *agr, struct casewriter *output) case MAX | FSTRING: case MIN | FSTRING: if (i->int1) - memcpy (v->s, i->string, var_get_width (i->dest)); + memcpy (value_str_rw (v, width), i->string, width); else - memset (v->s, ' ', var_get_width (i->dest)); + value_set_missing (v, width); break; case FGT: case FGT | FSTRING: @@ -1045,9 +1040,9 @@ dump_aggregate_info (struct agr_proc *agr, struct casewriter *output) case FIRST | FSTRING: case LAST | FSTRING: if (i->int1) - memcpy (v->s, i->string, var_get_width (i->dest)); + memcpy (value_str_rw (v, width), i->string, width); else - memset (v->s, ' ', var_get_width (i->dest)); + value_set_missing (v, width); break; case N_NO_VARS: v->f = i->dbl[0]; @@ -1069,7 +1064,7 @@ dump_aggregate_info (struct agr_proc *agr, struct casewriter *output) } } - casewriter_write (output, &c); + casewriter_write (output, c); } /* Resets the state for all the aggregate functions. */ @@ -1078,8 +1073,8 @@ initialize_aggregate_info (struct agr_proc *agr, const struct ccase *input) { struct agr_var *iter; - case_destroy (&agr->break_case); - case_clone (&agr->break_case, input); + case_unref (agr->break_case); + agr->break_case = case_ref (input); for (iter = agr->agr_vars; iter; iter = iter->next) { @@ -1102,7 +1097,12 @@ initialize_aggregate_info (struct agr_proc *agr, const struct ccase *input) break; case MEDIAN: { - struct case_ordering *ordering = case_ordering_create (); + struct caseproto *proto; + struct subcase ordering; + + proto = caseproto_create (); + proto = caseproto_add_width (proto, 0); + proto = caseproto_add_width (proto, 0); if ( ! iter->subject) iter->subject = var_create_internal (0); @@ -1110,9 +1110,11 @@ initialize_aggregate_info (struct agr_proc *agr, const struct ccase *input) if ( ! iter->weight) iter->weight = var_create_internal (1); - case_ordering_add_var (ordering, iter->subject, SRT_ASCEND); + subcase_init_var (&ordering, iter->subject, SC_ASCEND); + iter->writer = sort_create_writer (&ordering, proto); + subcase_destroy (&ordering); + caseproto_unref (proto); - iter->writer = sort_create_writer (ordering, 2); iter->cc = 0; } break;