X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Faggregate.c;h=eace3479a18b17568430c7c525293faf4c31a38e;hb=9e24f9a5c52650da0700066e8d9e72261c6492ad;hp=7430b819e90a2889aa4cb57be6575b180293335a;hpb=0eac9417b290d8373839cf72011fc86f5193741a;p=pspp-builds.git diff --git a/src/language/stats/aggregate.c b/src/language/stats/aggregate.c index 7430b819..eace3479 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, 2009, 2010 Free Software Foundation, Inc. + Copyright (C) 1997-9, 2000, 2006, 2008, 2009, 2010, 2011 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 @@ -16,40 +16,41 @@ #include +#include "language/stats/aggregate.h" + #include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include - -#include "aggregate.h" - -#include "minmax.h" -#include "xalloc.h" +#include "data/any-writer.h" +#include "data/case.h" +#include "data/casegrouper.h" +#include "data/casereader.h" +#include "data/casewriter.h" +#include "data/dataset.h" +#include "data/dictionary.h" +#include "data/file-handle-def.h" +#include "data/format.h" +#include "data/settings.h" +#include "data/subcase.h" +#include "data/sys-file-writer.h" +#include "data/variable.h" +#include "language/command.h" +#include "language/data-io/file-handle.h" +#include "language/lexer/lexer.h" +#include "language/lexer/variable-parser.h" +#include "language/stats/sort-criteria.h" +#include "libpspp/assertion.h" +#include "libpspp/i18n.h" +#include "libpspp/message.h" +#include "libpspp/misc.h" +#include "libpspp/pool.h" +#include "libpspp/str.h" +#include "math/moments.h" +#include "math/percentiles.h" +#include "math/sort.h" +#include "math/statistic.h" + +#include "gl/minmax.h" +#include "gl/xalloc.h" #include "gettext.h" #define _(msgid) gettext (msgid) @@ -87,41 +88,31 @@ struct agr_var struct casewriter *writer; }; -/* Aggregation functions. */ -enum - { - SUM, MEAN, MEDIAN, SD, MAX, MIN, PGT, PLT, PIN, POUT, FGT, FLT, FIN, - FOUT, N, NU, NMISS, NUMISS, FIRST, LAST, - - FUNC = 0x1f, /* Function mask. */ - FSTRING = 1<<5, /* String function bit. */ - }; - /* Attributes of aggregation functions. */ const struct agr_func agr_func_tab[] = { - {"SUM", N_("Sum of values"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, - {"MEAN", N_("Mean average"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, - {"MEDIAN", N_("Median average"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, - {"SD", N_("Standard deviation"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, - {"MAX", N_("Maximum value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, - {"MIN", N_("Minimum value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, - {"PGT", N_("Percentage greater than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 1}}, - {"PLT", N_("Percentage less than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 1}}, - {"PIN", N_("Percentage included in range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 1}}, - {"POUT", N_("Percentage excluded from range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 1}}, - {"FGT", N_("Fraction greater than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 3}}, - {"FLT", N_("Fraction less than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 3}}, - {"FIN", N_("Fraction included in range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 3}}, - {"FOUT", N_("Fraction excluded from range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 3}}, - {"N", N_("Number of cases"), AGR_SV_NO, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, - {"NU", N_("Number of cases (unweighted)"), AGR_SV_OPT, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, - {"NMISS", N_("Number of missing values"), AGR_SV_YES, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, + {"SUM", N_("Sum of values"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, + {"MEAN", N_("Mean average"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, + {"MEDIAN", N_("Median average"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, + {"SD", N_("Standard deviation"), AGR_SV_YES, 0, -1, {FMT_F, 8, 2}}, + {"MAX", N_("Maximum value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, + {"MIN", N_("Minimum value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, + {"PGT", N_("Percentage greater than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 1}}, + {"PLT", N_("Percentage less than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 1}}, + {"PIN", N_("Percentage included in range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 1}}, + {"POUT", N_("Percentage excluded from range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 1}}, + {"FGT", N_("Fraction greater than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 3}}, + {"FLT", N_("Fraction less than"), AGR_SV_YES, 1, VAL_NUMERIC, {FMT_F, 5, 3}}, + {"FIN", N_("Fraction included in range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 3}}, + {"FOUT", N_("Fraction excluded from range"), AGR_SV_YES, 2, VAL_NUMERIC, {FMT_F, 5, 3}}, + {"N", N_("Number of cases"), AGR_SV_NO, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, + {"NU", N_("Number of cases (unweighted)"), AGR_SV_OPT, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, + {"NMISS", N_("Number of missing values"), AGR_SV_YES, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, {"NUMISS", N_("Number of missing values (unweighted)"), AGR_SV_YES, 0, VAL_NUMERIC, {FMT_F, 7, 0}}, - {"FIRST", N_("First non-missing value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, - {"LAST", N_("Last non-missing value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, - {NULL, NULL, AGR_SV_NO, 0, -1, {-1, -1, -1}}, + {"FIRST", N_("First non-missing value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, + {"LAST", N_("Last non-missing value"), AGR_SV_YES, 0, VAL_STRING, {-1, -1, -1}}, + {NULL, NULL, AGR_SV_NO, 0, -1, {-1, -1, -1}}, }; /* Missing value types. */ @@ -185,11 +176,11 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) subcase_init_empty (&agr.sort); /* OUTFILE subcommand must be first. */ - lex_match (lexer, '/'); + lex_match (lexer, T_SLASH); if (!lex_force_match_id (lexer, "OUTFILE")) goto error; - lex_match (lexer, '='); - if (!lex_match (lexer, '*')) + lex_match (lexer, T_EQUALS); + if (!lex_match (lexer, T_ASTERISK)) { out_file = fh_parse (lexer, FH_REF_FILE | FH_REF_SCRATCH); if (out_file == NULL) @@ -198,7 +189,7 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) if (out_file == NULL && lex_match_id (lexer, "MODE")) { - lex_match (lexer, '='); + lex_match (lexer, T_EQUALS); if (lex_match_id (lexer, "ADDVARIABLES")) { agr.add_variables = true; @@ -225,14 +216,14 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) /* Read most of the subcommands. */ for (;;) { - lex_match (lexer, '/'); + lex_match (lexer, T_SLASH); if (lex_match_id (lexer, "MISSING")) { - lex_match (lexer, '='); + lex_match (lexer, T_EQUALS); if (!lex_match_id (lexer, "COLUMNWISE")) { - lex_error (lexer, _("while expecting COLUMNWISE")); + lex_error (lexer, _("expecting %s"), "COLUMNWISE"); goto error; } agr.missing = COLUMNWISE; @@ -241,11 +232,11 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) copy_documents = true; else if (lex_match_id (lexer, "PRESORTED")) presorted = true; - else if (lex_match_id (lexer, "BREAK")) + else if (lex_force_match_id (lexer, "BREAK")) { int i; - lex_match (lexer, '='); + lex_match (lexer, T_EQUALS); if (!parse_sort_criteria (lexer, dict, &agr.sort, &agr.break_vars, &saw_direction)) goto error; @@ -259,10 +250,8 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) break; } else - { - lex_error (lexer, _("expecting BREAK")); - goto error; - } + goto error; + } if (presorted && saw_direction) msg (SW, _("When PRESORTED is specified, specifying sorting directions " @@ -270,7 +259,7 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) "the same way as the input data.")); /* Read in the aggregate functions. */ - lex_match (lexer, '/'); + lex_match (lexer, T_SLASH); if (!parse_aggregate_functions (lexer, dict, &agr)) goto error; @@ -362,7 +351,8 @@ cmd_aggregate (struct lexer *lexer, struct dataset *ds) if (next_input == NULL) goto error; - proc_set_active_file (ds, next_input, agr.dict); + dataset_set_dict (ds, agr.dict); + dataset_set_source (ds, next_input); agr.dict = NULL; } else @@ -424,11 +414,11 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, ds_init_empty (&function_name); /* Parse the list of target variables. */ - while (!lex_match (lexer, '=')) + while (!lex_match (lexer, T_EQUALS)) { size_t n_dest_prev = n_dest; - if (!parse_DATA_LIST_vars (lexer, &dest, &n_dest, + if (!parse_DATA_LIST_vars (lexer, dict, &dest, &n_dest, (PV_APPEND | PV_SINGLE | PV_NO_SCRATCH | PV_NO_DUPLICATE))) goto error; @@ -444,15 +434,10 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, - if (lex_token (lexer) == T_STRING) + if (lex_is_string (lexer)) { - struct string label; - ds_init_string (&label, lex_tokstr (lexer)); - - ds_truncate (&label, 255); - dest_label[n_dest - 1] = ds_xstrdup (&label); + dest_label[n_dest - 1] = xstrdup (lex_tokcstr (lexer)); lex_get (lexer); - ds_destroy (&label); } } @@ -463,14 +448,8 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, goto error; } - exclude = MV_ANY; - - ds_assign_string (&function_name, lex_tokstr (lexer)); - - ds_chomp (&function_name, '.'); - - if (lex_tokid(lexer)[strlen (lex_tokid (lexer)) - 1] == '.') - exclude = MV_SYSTEM; + ds_assign_substring (&function_name, lex_tokss (lexer)); + exclude = ds_chomp_byte (&function_name, '.') ? MV_SYSTEM : MV_ANY; for (function = agr_func_tab; function->name; function++) if (!strcasecmp (function->name, ds_cstr (&function_name))) @@ -486,11 +465,11 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, lex_get (lexer); /* Check for leading lparen. */ - if (!lex_match (lexer, '(')) + if (!lex_match (lexer, T_LPAREN)) { if (function->src_vars == AGR_SV_YES) { - lex_error (lexer, _("expecting `('")); + lex_force_match (lexer, T_LPAREN); goto error; } } @@ -516,10 +495,12 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, { int type; - lex_match (lexer, ','); - if (lex_token (lexer) == T_STRING) + lex_match (lexer, T_COMMA); + if (lex_is_string (lexer)) { - arg[i].c = ds_xstrdup (lex_tokstr (lexer)); + arg[i].c = recode_string (dict_get_encoding (agr->dict), + "UTF-8", lex_tokcstr (lexer), + -1); type = VAL_STRING; } else if (lex_is_number (lexer)) @@ -546,11 +527,8 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, } /* Trailing rparen. */ - if (!lex_match (lexer, ')')) - { - lex_error (lexer, _("expecting `)'")); - goto error; - } + if (!lex_force_match (lexer, T_RPAREN)) + goto error; /* Now check that the number of source variables match the number of target variables. If we check earlier @@ -637,12 +615,16 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, struct fmt_spec f; v->src = NULL; destvar = dict_create_var (agr->dict, dest[i], 0); - if (func_index == N && dict_get_weight (dict) != NULL) - f = fmt_for_output (FMT_F, 8, 2); - else - f = function->format; - var_set_both_formats (destvar, &f); - } + if (destvar != NULL) + { + if ((func_index == N || func_index == NMISS) + && dict_get_weight (dict) != NULL) + f = fmt_for_output (FMT_F, 8, 2); + else + f = function->format; + var_set_both_formats (destvar, &f); + } + } if (!destvar) { @@ -656,7 +638,8 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, free (dest[i]); if (dest_label[i]) - var_set_label (destvar, dest_label[i]); + var_set_label (destvar, dest_label[i], + dict_get_encoding (agr->dict), true); v->dest = destvar; } @@ -687,9 +670,9 @@ parse_aggregate_functions (struct lexer *lexer, const struct dictionary *dict, free (dest); free (dest_label); - if (!lex_match (lexer, '/')) + if (!lex_match (lexer, T_SLASH)) { - if (lex_token (lexer) == '.') + if (lex_token (lexer) == T_ENDCMD) return true; lex_error (lexer, "expecting end of command"); @@ -827,6 +810,7 @@ accumulate_aggregate_info (struct agr_proc *agr, const struct ccase *input) iter->int1 = 1; break; case MAX | FSTRING: + /* Need to do some kind of Unicode collation thingy here */ 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; @@ -1003,21 +987,22 @@ dump_aggregate_info (const struct agr_proc *agr, struct casewriter *output, cons break; case MEDIAN: { - struct casereader *sorted_reader; - struct percentile *median = percentile_create (0.5, i->cc); - struct order_stats *os = &median->parent; - - sorted_reader = casewriter_make_reader (i->writer); - - order_stats_accumulate (&os, 1, - sorted_reader, - i->weight, - i->subject, - i->exclude); - - v->f = percentile_calculate (median, PC_HAVERAGE); - - statistic_destroy (&median->parent.parent); + if ( i->writer) + { + struct percentile *median = percentile_create (0.5, i->cc); + struct order_stats *os = &median->parent; + struct casereader *sorted_reader = casewriter_make_reader (i->writer); + i->writer = NULL; + + order_stats_accumulate (&os, 1, + sorted_reader, + i->weight, + i->subject, + i->exclude); + i->dbl[0] = percentile_calculate (median, PC_HAVERAGE); + statistic_destroy (&median->parent.parent); + } + v->f = i->dbl[0]; } break; case SD: