/* PSPP - a program for statistical analysis.
- Copyright (C) 2007, 2008, 2009, 2010, 2011 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
#include <stdlib.h>
-#include <data/gnumeric-reader.h>
-#include <data/psql-reader.h>
-
-#include <data/dictionary.h>
-#include <data/format.h>
-#include <data/procedure.h>
-#include <data/settings.h>
-#include <language/command.h>
-#include <language/data-io/data-parser.h>
-#include <language/data-io/data-reader.h>
-#include <language/data-io/file-handle.h>
-#include <language/data-io/placement-parser.h>
-#include <language/lexer/format-parser.h>
-#include <language/lexer/lexer.h>
-#include <libpspp/message.h>
-
-#include "xalloc.h"
+#include <string.h>
+
+#include "data/dataset.h"
+#include "data/dictionary.h"
+#include "data/format.h"
+#include "data/gnumeric-reader.h"
+#include "data/ods-reader.h"
+#include "data/spreadsheet-reader.h"
+#include "data/psql-reader.h"
+#include "data/settings.h"
+#include "language/command.h"
+#include "language/data-io/data-parser.h"
+#include "language/data-io/data-reader.h"
+#include "language/data-io/file-handle.h"
+#include "language/data-io/placement-parser.h"
+#include "language/lexer/format-parser.h"
+#include "language/lexer/lexer.h"
+#include "libpspp/cast.h"
+#include "libpspp/i18n.h"
+#include "libpspp/message.h"
+
+#include "gl/xalloc.h"
#include "gettext.h"
#define _(msgid) gettext (msgid)
#define N_(msgid) (msgid)
-static int parse_get_gnm (struct lexer *lexer, struct dataset *);
+static bool parse_spreadsheet (struct lexer *lexer, char **filename,
+ struct spreadsheet_read_options *opts);
+
+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 *);
int
cmd_get_data (struct lexer *lexer, struct dataset *ds)
{
- lex_force_match (lexer, T_SLASH);
+ char *tok = NULL;
+ 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;
- lex_force_match (lexer, T_EQUALS);
+ const char *s = lex_tokcstr (lexer);
- if (lex_match_id (lexer, "GNM"))
- return parse_get_gnm (lexer, ds);
- else if (lex_match_id (lexer, "TXT"))
- return parse_get_txt (lexer, ds);
+ if (s)
+ tok = strdup (s);
+
+ if (lex_match_id (lexer, "TXT"))
+ {
+ free (tok);
+ return parse_get_txt (lexer, ds);
+ }
else if (lex_match_id (lexer, "PSQL"))
- return parse_get_psql (lexer, ds);
+ {
+ free (tok);
+ return parse_get_psql (lexer, ds);
+ }
+ else if (lex_match_id (lexer, "GNM") ||
+ lex_match_id (lexer, "ODS"))
+ {
+ char *filename = NULL;
+ if (!parse_spreadsheet (lexer, &filename, &opts))
+ goto error;
+
+ struct spreadsheet *spreadsheet = NULL;
+ if ( 0 == strncasecmp (tok, "GNM", 3))
+ 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_clone (spreadsheet->dict));
+ dataset_set_source (ds, reader);
+ free (tok);
+ destroy_spreadsheet_read_info (&opts);
+ spreadsheet_unref (spreadsheet);
+ return CMD_SUCCESS;
+ }
+ spreadsheet_unref (spreadsheet);
+ }
+ else
+ msg (SE, _("Unsupported TYPE %s."), tok);
+
- msg (SE, _("Unsupported TYPE %s."), lex_tokcstr (lexer));
+ error:
+ destroy_spreadsheet_read_info (&opts);
+ free (tok);
return CMD_FAILURE;
}
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;
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"))
{
struct casereader *reader = psql_open_reader (&psql, &dict);
if ( reader )
- proc_set_active_file (ds, reader, dict);
+ {
+ dataset_set_dict (ds, dict);
+ dataset_set_source (ds, reader);
+ }
}
ds_destroy (&psql.sql);
return CMD_FAILURE;
}
-static int
-parse_get_gnm (struct lexer *lexer, struct dataset *ds)
+static bool
+parse_spreadsheet (struct lexer *lexer, char **filename,
+ struct spreadsheet_read_options *opts)
{
- struct gnumeric_read_info gri = {NULL, NULL, NULL, 1, true, -1};
+ opts->sheet_index = 1;
+ opts->sheet_name = NULL;
+ opts->cell_range = 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;
- gri.file_name = ss_xstrdup (lex_tokss (lexer));
+ *filename = utf8_to_filename (lex_tokcstr (lexer));
lex_get (lexer);
if ( lex_match_id (lexer, "ASSUMEDSTRWIDTH"))
{
lex_match (lexer, T_EQUALS);
- gri.asw = lex_integer (lexer);
+ if (lex_force_int (lexer))
+ {
+ opts->asw = lex_integer (lexer);
+ lex_get (lexer);
+ }
}
else if (lex_match_id (lexer, "SHEET"))
{
if ( ! lex_force_string (lexer) )
goto error;
- gri.sheet_name = ss_xstrdup (lex_tokss (lexer));
- gri.sheet_index = -1;
+ opts->sheet_name = ss_xstrdup (lex_tokss (lexer));
+ opts->sheet_index = -1;
+
+ lex_get (lexer);
}
else if (lex_match_id (lexer, "INDEX"))
{
- gri.sheet_index = lex_integer (lexer);
+ if (!lex_force_int (lexer))
+ goto error;
+
+ opts->sheet_index = lex_integer (lexer);
+ if (opts->sheet_index <= 0)
+ {
+ msg (SE, _("The sheet index must be greater than or equal to 1"));
+ goto error;
+ }
+ lex_get (lexer);
}
else
- goto error;
+ {
+ msg (SE, _("%s must be followed by either \"%s\" or \"%s\"."),
+ "/SHEET", "NAME", "INDEX");
+ goto error;
+ }
}
else if (lex_match_id (lexer, "CELLRANGE"))
{
if (lex_match_id (lexer, "FULL"))
{
- gri.cell_range = NULL;
- lex_put_back (lexer, T_ID);
+ opts->cell_range = NULL;
}
else if (lex_match_id (lexer, "RANGE"))
{
if ( ! lex_force_string (lexer) )
goto error;
- gri.cell_range = ss_xstrdup (lex_tokss (lexer));
+ opts->cell_range = ss_xstrdup (lex_tokss (lexer));
+ lex_get (lexer);
}
else
- goto error;
+ {
+ msg (SE, _("%s must be followed by either \"%s\" or \"%s\"."),
+ "/CELLRANGE", "FULL", "RANGE");
+ goto error;
+ }
}
else if (lex_match_id (lexer, "READNAMES"))
{
if ( lex_match_id (lexer, "ON"))
{
- gri.read_names = true;
+ opts->read_names = true;
}
else if (lex_match_id (lexer, "OFF"))
{
- gri.read_names = false;
+ opts->read_names = false;
}
else
- goto error;
- lex_put_back (lexer, T_ID);
+ {
+ msg (SE, _("%s must be followed by either \"%s\" or \"%s\"."),
+ "/READNAMES", "ON", "OFF");
+ goto error;
+ }
}
else
{
- printf ("Unknown data file type `%s'\n", lex_tokcstr (lexer));
+ lex_error (lexer, NULL);
goto error;
}
- lex_get (lexer);
}
- {
- struct dictionary *dict = NULL;
- struct casereader *reader = gnumeric_open_reader (&gri, &dict);
-
- if ( reader )
- proc_set_active_file (ds, reader, dict);
- }
-
- free (gri.file_name);
- free (gri.sheet_name);
- free (gri.cell_range);
- return CMD_SUCCESS;
+ return true;
error:
-
- free (gri.file_name);
- free (gri.sheet_name);
- free (gri.cell_range);
- return CMD_FAILURE;
+ return false;
}
+
static bool
set_type (struct data_parser *parser, const char *subcommand,
enum data_parser_type type, bool *has_type)
parse_get_txt (struct lexer *lexer, struct dataset *ds)
{
struct data_parser *parser = NULL;
- struct dictionary *dict = dict_create ();
+ struct dictionary *dict = dict_create (get_default_encoding ());
struct file_handle *fh = NULL;
struct dfm_reader *reader = NULL;
+ char *encoding = NULL;
char *name = NULL;
int record;
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);
- fh = fh_parse (lexer, FH_REF_FILE | FH_REF_INLINE);
+ 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;
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 (;;)
if (!lex_force_match (lexer, T_SLASH))
goto error;
- if (lex_match_id (lexer, "ARRANGEMENT"))
+ if (lex_match_id (lexer, "ENCODING"))
+ {
+ lex_match (lexer, T_EQUALS);
+ if (!lex_force_string (lexer))
+ goto error;
+
+ free (encoding);
+ encoding = ss_xstrdup (lex_tokss (lexer));
+
+ lex_get (lexer);
+ }
+ else if (lex_match_id (lexer, "ARRANGEMENT"))
{
bool ok;
DP_DELIMITED, &has_type);
else
{
- lex_error (lexer, _("expecting %s or %s"), "FIXED", "DELIMITED");
+ lex_error_expecting (lexer, "FIXED", "DELIMITED", NULL_SENTINEL);
goto error;
}
if (!ok)
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);
}
else
{
- lex_error (lexer, _("expecting %s or %s"), "LINE", "VARIABLES");
+ lex_error_expecting (lexer, "LINE", "VARIABLES", NULL_SENTINEL);
goto error;
}
}
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));
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))
{
if (!lex_force_string (lexer))
goto error;
+ /* XXX should support multibyte UTF-8 characters */
s = lex_tokss (lexer);
if (ss_match_string (&s, ss_cstr ("\\t")))
ds_put_cstr (&hard_seps, "\t");
if (!lex_force_string (lexer))
goto error;
+ /* XXX should support multibyte UTF-8 characters */
if (settings_get_syntax () == COMPATIBLE
&& ss_length (lex_tokss (lexer)) != 1)
{
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
{
- lex_error (lexer, _("expecting %s"), "VARIABLES");
+ lex_error_expecting (lexer, "VARIABLES", NULL_SENTINEL);
goto error;
}
}
lex_get (lexer);
}
- if (!lex_force_id (lexer))
- goto error;
- name = xstrdup (lex_tokcstr (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, 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
{
+ char fmt_type_name[FMT_TYPE_LEN_MAX + 1];
+ enum fmt_type fmt_type;
+ int w, d;
+
if (!parse_column_range (lexer, 0, &fc, &lc, NULL))
goto error;
- if (!parse_format_specifier_name (lexer, &input.type))
+
+ /* Accept a format (e.g. F8.2) or just a type name (e.g. DOLLAR). */
+ if (!parse_abstract_format_specifier (lexer, fmt_type_name, &w, &d))
goto error;
+ if (!fmt_from_name (fmt_type_name, &fmt_type))
+ {
+ 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)
+ {
+ output.type = fmt_type;
+ output.w = w;
+ output.d = d;
+ if (!fmt_check_output (&output))
+ goto error;
+ }
+ else
+ output = fmt_for_output_from_input (&input);
}
- output = fmt_for_output_from_input (&input);
-
v = dict_create_var (dict, name, fmt_var_width (&input));
if (v == NULL)
{
goto error;
}
var_set_both_formats (v, &output);
-
if (type == DP_DELIMITED)
data_parser_add_delimited_field (parser, &input,
var_get_case_index (v),
}
while (lex_token (lexer) != T_ENDCMD);
- reader = dfm_open_reader (fh, lexer);
+ reader = dfm_open_reader (fh, lexer, encoding);
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;
error:
dict_destroy (dict);
fh_unref (fh);
free (name);
+ free (encoding);
return CMD_CASCADING_FAILURE;
}
+
+
+static void
+destroy_spreadsheet_read_info (struct spreadsheet_read_options *opts)
+{
+ free (opts->cell_range);
+ free (opts->sheet_name);
+}