X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;ds=sidebyside;f=src%2Flanguage%2Fdata-io%2Fget-data.c;h=dff449851bd91ff8a5bc3b3b18a26b196242f603;hb=5e328ed3392221e6c201a60e6d4e6638fe7f32de;hp=cca86a2bd838bf5b83f9f89f9e0530a83301a188;hpb=5c210c96542db2a357cf1aa6b690ef2ea88a54b1;p=pspp diff --git a/src/language/data-io/get-data.c b/src/language/data-io/get-data.c index cca86a2bd8..dff449851b 100644 --- a/src/language/data-io/get-data.c +++ b/src/language/data-io/get-data.c @@ -1,5 +1,6 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 2007, 2008, 2009, 2010, 2011, 2012, 2013 Free Software Foundation, Inc. + Copyright (C) 2007, 2008, 2009, 2010, 2011, 2012, + 2013, 2015, 2016 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 @@ -45,10 +46,10 @@ #define _(msgid) gettext (msgid) #define N_(msgid) (msgid) -static bool parse_spreadsheet (struct lexer *lexer, struct spreadsheet_read_info *sri, +static bool parse_spreadsheet (struct lexer *lexer, char **filename, struct spreadsheet_read_options *opts); -static void destroy_spreadsheet_read_info (struct spreadsheet_read_info *, struct spreadsheet_read_options *); +static void destroy_spreadsheet_read_info (struct spreadsheet_read_options *); static int parse_get_txt (struct lexer *lexer, struct dataset *); static int parse_get_psql (struct lexer *lexer, struct dataset *); @@ -57,14 +58,28 @@ int cmd_get_data (struct lexer *lexer, struct dataset *ds) { char *tok = NULL; - lex_force_match (lexer, T_SLASH); + struct spreadsheet_read_options opts; + + opts.sheet_name = NULL; + opts.sheet_index = -1; + opts.cell_range = NULL; + opts.read_names = false; + opts.asw = -1; + + if (! lex_force_match (lexer, T_SLASH)) + goto error; if (!lex_force_match_id (lexer, "TYPE")) - return CMD_FAILURE; + goto error; + + if (!lex_force_match (lexer, T_EQUALS)) + goto error; + + const char *s = lex_tokcstr (lexer); - lex_force_match (lexer, T_EQUALS); + if (s) + tok = strdup (s); - tok = strdup (lex_tokcstr (lexer)); if (lex_match_id (lexer, "TXT")) { free (tok); @@ -75,43 +90,45 @@ cmd_get_data (struct lexer *lexer, struct dataset *ds) free (tok); return parse_get_psql (lexer, ds); } - else if (lex_match_id (lexer, "GNM") || + else if (lex_match_id (lexer, "GNM") || lex_match_id (lexer, "ODS")) { - struct casereader *reader = NULL; - struct dictionary *dict = NULL; - struct spreadsheet_read_info sri; - struct spreadsheet_read_options opts; - if (!parse_spreadsheet (lexer, &sri, &opts)) + char *filename = NULL; + if (!parse_spreadsheet (lexer, &filename, &opts)) goto error; + struct spreadsheet *spreadsheet = NULL; if ( 0 == strncasecmp (tok, "GNM", 3)) - { - struct spreadsheet *spreadsheet = gnumeric_probe (sri.file_name); - reader = gnumeric_make_reader (spreadsheet, &sri, &opts); - dict = spreadsheet->dict; - } - else if (0 == strncasecmp (tok, "ODS", 3)) - { - struct spreadsheet *spreadsheet = ods_probe (sri.file_name); - reader = ods_make_reader (spreadsheet, &sri, &opts); - dict = spreadsheet->dict; - } + spreadsheet = gnumeric_probe (filename, true); + else if ( 0 == strncasecmp (tok, "ODS", 3)) + spreadsheet = ods_probe (filename, true); + + if (spreadsheet == NULL) + { + msg (SE, _("error reading file `%s'"), filename); + free (filename); + goto error; + } + free (filename); + struct casereader *reader = spreadsheet_make_reader (spreadsheet, &opts); if (reader) { - dataset_set_dict (ds, dict); + dataset_set_dict (ds, dict_clone (spreadsheet->dict)); dataset_set_source (ds, reader); - destroy_spreadsheet_read_info (&sri, &opts); free (tok); + destroy_spreadsheet_read_info (&opts); + spreadsheet_unref (spreadsheet); return CMD_SUCCESS; } - destroy_spreadsheet_read_info (&sri, &opts); + spreadsheet_unref (spreadsheet); } else msg (SE, _("Unsupported TYPE %s."), tok); + error: + destroy_spreadsheet_read_info (&opts); free (tok); return CMD_FAILURE; } @@ -126,12 +143,14 @@ parse_get_psql (struct lexer *lexer, struct dataset *ds) psql.bsize = -1; ds_init_empty (&psql.sql); - lex_force_match (lexer, T_SLASH); + if (! lex_force_match (lexer, T_SLASH)) + goto error; if (!lex_force_match_id (lexer, "CONNECT")) goto error; - lex_force_match (lexer, T_EQUALS); + if (! lex_force_match (lexer, T_EQUALS)) + goto error; if (!lex_force_string (lexer)) goto error; @@ -145,14 +164,20 @@ parse_get_psql (struct lexer *lexer, struct dataset *ds) if ( lex_match_id (lexer, "ASSUMEDSTRWIDTH")) { lex_match (lexer, T_EQUALS); - psql.str_width = lex_integer (lexer); - lex_get (lexer); + if (lex_force_int (lexer)) + { + psql.str_width = lex_integer (lexer); + lex_get (lexer); + } } else if ( lex_match_id (lexer, "BSIZE")) { lex_match (lexer, T_EQUALS); - psql.bsize = lex_integer (lexer); - lex_get (lexer); + if (lex_force_int (lexer)) + { + psql.bsize = lex_integer (lexer); + lex_get (lexer); + } } else if ( lex_match_id (lexer, "UNENCRYPTED")) { @@ -193,27 +218,28 @@ parse_get_psql (struct lexer *lexer, struct dataset *ds) } static bool -parse_spreadsheet (struct lexer *lexer, struct spreadsheet_read_info *sri, +parse_spreadsheet (struct lexer *lexer, char **filename, struct spreadsheet_read_options *opts) { opts->sheet_index = 1; opts->sheet_name = NULL; opts->cell_range = NULL; - sri->read_names = true; - sri->asw = -1; - sri->file_name = NULL; + opts->read_names = true; + opts->asw = -1; - lex_force_match (lexer, T_SLASH); + if (! lex_force_match (lexer, T_SLASH)) + goto error; if (!lex_force_match_id (lexer, "FILE")) goto error; - lex_force_match (lexer, T_EQUALS); + if (! lex_force_match (lexer, T_EQUALS)) + goto error; if (!lex_force_string (lexer)) goto error; - sri->file_name = utf8_to_filename (lex_tokcstr (lexer)); + *filename = utf8_to_filename (lex_tokcstr (lexer)); lex_get (lexer); @@ -222,8 +248,11 @@ parse_spreadsheet (struct lexer *lexer, struct spreadsheet_read_info *sri, if ( lex_match_id (lexer, "ASSUMEDSTRWIDTH")) { lex_match (lexer, T_EQUALS); - sri->asw = lex_integer (lexer); - lex_get (lexer); + if (lex_force_int (lexer)) + { + opts->asw = lex_integer (lexer); + lex_get (lexer); + } } else if (lex_match_id (lexer, "SHEET")) { @@ -240,6 +269,9 @@ parse_spreadsheet (struct lexer *lexer, struct spreadsheet_read_info *sri, } else if (lex_match_id (lexer, "INDEX")) { + if (!lex_force_int (lexer)) + goto error; + opts->sheet_index = lex_integer (lexer); if (opts->sheet_index <= 0) { @@ -284,11 +316,11 @@ parse_spreadsheet (struct lexer *lexer, struct spreadsheet_read_info *sri, if ( lex_match_id (lexer, "ON")) { - sri->read_names = true; + opts->read_names = true; } else if (lex_match_id (lexer, "OFF")) { - sri->read_names = false; + opts->read_names = false; } else { @@ -307,7 +339,6 @@ parse_spreadsheet (struct lexer *lexer, struct spreadsheet_read_info *sri, return true; error: - destroy_spreadsheet_read_info (sri, opts); return false; } @@ -347,11 +378,13 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) enum data_parser_type type; bool has_type; - lex_force_match (lexer, T_SLASH); + if (! lex_force_match (lexer, T_SLASH)) + goto error; if (!lex_force_match_id (lexer, "FILE")) goto error; - lex_force_match (lexer, T_EQUALS); + if (! lex_force_match (lexer, T_EQUALS)) + goto error; fh = fh_parse (lexer, FH_REF_FILE | FH_REF_INLINE, NULL); if (fh == NULL) goto error; @@ -361,6 +394,7 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) data_parser_set_type (parser, DP_DELIMITED); data_parser_set_span (parser, false); data_parser_set_quotes (parser, ss_empty ()); + data_parser_set_quote_escape (parser, true); data_parser_set_empty_line_has_field (parser, true); for (;;) @@ -404,7 +438,7 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) goto error; if (lex_integer (lexer) < 1) { - msg (SE, _("Value of FIRSTCASE must be 1 or greater.")); + msg (SE, _("Value of %s must be 1 or greater."), "FIRSTCASE"); goto error; } data_parser_set_skip (parser, lex_integer (lexer) - 1); @@ -442,7 +476,7 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) goto error; if (lex_integer (lexer) < 1) { - msg (SE, _("Value of FIXCASE must be at least 1.")); + msg (SE, _("Value of %s must be 1 or greater."), "FIXCASE"); goto error; } data_parser_set_records (parser, lex_integer (lexer)); @@ -453,33 +487,22 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) lex_match (lexer, T_EQUALS); if (lex_match (lexer, T_ALL)) { - data_parser_set_case_limit (parser, -1); - data_parser_set_case_percent (parser, 100); + /* Nothing to do. */ } else if (lex_match_id (lexer, "FIRST")) { if (!lex_force_int (lexer)) goto error; - if (lex_integer (lexer) < 1) - { - msg (SE, _("Value of FIRST must be at least 1.")); - goto error; - } - data_parser_set_case_limit (parser, lex_integer (lexer)); lex_get (lexer); } else if (lex_match_id (lexer, "PERCENT")) { if (!lex_force_int (lexer)) goto error; - if (lex_integer (lexer) < 1 || lex_integer (lexer) > 100) - { - msg (SE, _("Value of PERCENT must be between 1 and 100.")); - goto error; - } - data_parser_set_case_percent (parser, lex_integer (lexer)); lex_get (lexer); } + msg (SW, _("Ignoring obsolete IMPORTCASES subcommand. (N OF CASES " + "or SAMPLE may be used to substitute.)")); } else if (lex_match_id_n (lexer, "DELIMITERS", 4)) { @@ -533,9 +556,6 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) data_parser_set_quotes (parser, lex_tokss (lexer)); lex_get (lexer); } - else if (settings_get_syntax () == ENHANCED - && lex_match_id (lexer, "ESCAPE")) - data_parser_set_quote_escape (parser, true); else if (lex_match_id (lexer, "VARIABLES")) break; else @@ -579,18 +599,26 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) lex_get (lexer); } + const char * tstr = lex_tokcstr (lexer); + if (tstr == NULL) + { + lex_error (lexer, NULL); + goto error; + } + name = xstrdup (tstr); if (!lex_force_id (lexer) - || !dict_id_is_valid (dict, lex_tokcstr (lexer), true)) - goto error; - name = xstrdup (lex_tokcstr (lexer)); + || !dict_id_is_valid (dict, name, true)) + { + goto error; + } lex_get (lexer); - if (type == DP_DELIMITED) { if (!parse_format_specifier (lexer, &input) - || !fmt_check_input (&input)) - goto error; - + || !fmt_check_input (&input)) + { + goto error; + } output = fmt_for_output_from_input (&input); } else @@ -610,14 +638,12 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) msg (SE, _("Unknown format type `%s'."), fmt_type_name); goto error; } - /* Compose input format. */ input.type = fmt_type; input.w = lc - fc + 1; input.d = 0; if (!fmt_check_input (&input)) goto error; - /* Compose output format. */ if (w != 0) { @@ -630,7 +656,6 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) else output = fmt_for_output_from_input (&input); } - v = dict_create_var (dict, name, fmt_var_width (&input)); if (v == NULL) { @@ -638,7 +663,6 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) goto error; } var_set_both_formats (v, &output); - if (type == DP_DELIMITED) data_parser_add_delimited_field (parser, &input, var_get_case_index (v), @@ -655,7 +679,7 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) if (reader == NULL) goto error; - data_parser_make_active_file (parser, ds, reader, dict); + data_parser_make_active_file (parser, ds, reader, dict, NULL, NULL); fh_unref (fh); free (encoding); return CMD_SUCCESS; @@ -670,11 +694,9 @@ parse_get_txt (struct lexer *lexer, struct dataset *ds) } -static void -destroy_spreadsheet_read_info (struct spreadsheet_read_info *sri, - struct spreadsheet_read_options *opts) +static void +destroy_spreadsheet_read_info (struct spreadsheet_read_options *opts) { - free (opts->sheet_name); free (opts->cell_range); - free (sri->file_name); + free (opts->sheet_name); }