/* PSPP - a program for statistical analysis.
- Copyright (C) 2007 Free Software Foundation, Inc.
+ Copyright (C) 2007, 2009, 2010 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 <config.h>
#include <libpspp/message.h>
+#include <libpspp/misc.h>
+
+#include "minmax.h"
#include "gettext.h"
#define _(msgid) gettext (msgid)
#include <data/identifier.h>
#include <assert.h>
+/* Default width of string variables. */
+#define GNUMERIC_DEFAULT_WIDTH 8
static void gnm_file_casereader_destroy (struct casereader *, void *);
-static bool gnm_file_casereader_read (struct casereader *, void *,
- struct ccase *);
+static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
static const struct casereader_class gnm_file_casereader_class =
{
int stop_row;
int stop_col;
-
- size_t value_cnt;
+ struct caseproto *proto;
struct dictionary *dict;
- struct ccase first_case;
+ struct ccase *first_case;
bool used_first_case;
};
xmlFreeTextReader (r->xtr);
if ( ! r->used_first_case )
- case_destroy (&r->first_case);
+ case_unref (r->first_case);
+
+ caseproto_unref (r->proto);
free (r);
}
}
-
-/*
- Change SUGGESTION until it's a valid name that can be added to DICT.
-*/
-static void
-devise_name (const struct dictionary *dict, struct string *name, int *x)
-{
- struct string basename;
- if ( ds_is_empty (name))
- ds_init_cstr (&basename, "var");
- else
- ds_init_string (&basename, name);
- do
- {
- ds_clear (name);
- ds_put_format (name, "%s%d", ds_cstr (&basename), ++(*x));
- }
- while (NULL != dict_lookup_var (dict, ds_cstr (name)) );
-
- ds_destroy (&basename);
-}
-
-/*
- Mutate NAME of a variable, which is gauranteed to be valid for the
- dictionary DICT.
-*/
-static void
-munge_name (const struct dictionary *dict, struct string *name)
-{
- int x = 0;
-
- if (! ds_is_empty (name))
- {
- /* Change all the invalid characters to valid ones */
- char *s;
-
- s = ds_data (name);
-
- if ( !lex_is_id1 (*s))
- *s = '@';
-
- s++;
-
- while (s < ds_data (name) + ds_length (name))
- {
- if ( !lex_is_idn (*s))
- *s = '_';
- s++;
- }
-
- assert (var_is_valid_name (ds_cstr (name), false));
- }
-
- while (ds_is_empty (name) || NULL != dict_lookup_var (dict, ds_cstr (name)) )
- {
- devise_name (dict, name, &x);
- }
-}
-
-
/*
Sets the VAR of case C, to the value corresponding to the xml string XV
*/
convert_xml_string_to_value (struct ccase *c, const struct variable *var,
const xmlChar *xv)
{
- char *text;
- int n_bytes = 0;
union value *v = case_data_rw (c, var);
- text = recode_string (CONV_UTF8_TO_PSPP, (const char *) xv, -1);
-
- if ( text)
- n_bytes = MIN (var_get_width (var), strlen (text));
-
- if ( var_is_alpha (var))
- {
- memcpy (v->s, text, n_bytes);
- }
+ if (xv == NULL)
+ value_set_missing (v, var_get_width (var));
+ else if ( var_is_alpha (var))
+ value_copy_str_rpad (v, var_get_width (var), xv, ' ');
else
{
+ const char *text = (const char *) xv;
char *endptr;
+
errno = 0;
v->f = strtod (text, &endptr);
if ( errno != 0 || endptr == text)
v->f = SYSMIS;
}
-
- free (text);
}
struct var_spec
struct casereader *
gnumeric_open_reader (struct gnumeric_read_info *gri, struct dictionary **dict)
{
+ unsigned long int vstart = 0;
int ret;
casenumber n_cases = CASENUMBER_MAX;
int i;
if ( NULL == gz)
{
- msg (ME, _("Error opening \"%s\" for reading as a gnumeric file: %s."),
+ msg (ME, _("Error opening \"%s\" for reading as a Gnumeric file: %s."),
gri->file_name, strerror (errno));
goto error;
if ( idx >= n_var_specs )
{
n_var_specs = idx + 1 ;
- var_spec = realloc (var_spec, sizeof (*var_spec) * n_var_specs);
+ var_spec = xrealloc (var_spec, sizeof (*var_spec) * n_var_specs);
var_spec [idx].name = NULL;
var_spec [idx].width = -1;
var_spec [idx].first_value = NULL;
if ( r->node_type == XML_READER_TYPE_TEXT )
{
- char *text ;
xmlChar *value = xmlTextReaderValue (r->xtr);
-
- text = recode_string (CONV_UTF8_TO_PSPP, (const char *) value, -1);
+ const char *text = (const char *) value;
if ( r->row < r->start_row)
{
if ( gri->read_names )
{
- var_spec [idx].name = strdup (text);
+ var_spec [idx].name = xstrdup (text);
}
}
else
if (-1 == var_spec [idx].width )
var_spec [idx].width = (gri->asw == -1) ?
- ROUND_UP (strlen(text), MAX_SHORT_STRING) : gri->asw;
+ ROUND_UP (strlen(text), GNUMERIC_DEFAULT_WIDTH) : gri->asw;
}
free (value);
- free (text);
}
else if ( r->node_type == XML_READER_TYPE_ELEMENT
&& r->state == STATE_CELL)
/* Create the dictionary and populate it */
*dict = r->dict = dict_create ();
- r->value_cnt = 0;
+ dict_set_encoding (r->dict, (const char *) xmlTextReaderConstEncoding (r->xtr));
for (i = 0 ; i < n_var_specs ; ++i )
{
- struct string name;
+ char name[VAR_NAME_LEN + 1];
- /* Probably no data exists for this variable, so allocate a default width */
+ /* Probably no data exists for this variable, so allocate a
+ default width */
if ( var_spec[i].width == -1 )
- var_spec[i].width = MAX_SHORT_STRING;
-
- r->value_cnt += value_cnt_from_width (var_spec[i].width);
-
- if (var_spec[i].name)
- ds_init_cstr (&name, var_spec[i].name);
- else
- ds_init_empty (&name);
-
- munge_name (r->dict, &name);
+ var_spec[i].width = GNUMERIC_DEFAULT_WIDTH;
+ if ( ! dict_make_unique_var_name (r->dict, var_spec[i].name,
+ &vstart, name))
+ {
+ msg (ME, _("Cannot create variable name from %s"), var_spec[i].name);
+ goto error;
+ }
- dict_create_var (r->dict, ds_cstr (&name), var_spec[i].width);
-
- ds_destroy (&name);
+ dict_create_var (r->dict, name, var_spec[i].width);
}
/* Create the first case, and cache it */
goto error;
}
- case_create (&r->first_case, r->value_cnt);
- memset (case_data_rw_idx (&r->first_case, 0)->s,
- ' ', MAX_SHORT_STRING * r->value_cnt);
+ r->proto = caseproto_ref (dict_get_proto (r->dict));
+ r->first_case = case_create (r->proto);
+ case_set_missing (r->first_case);
for ( i = 0 ; i < n_var_specs ; ++i )
{
const struct variable *var = dict_get_var (r->dict, i);
- convert_xml_string_to_value (&r->first_case, var,
+ convert_xml_string_to_value (r->first_case, var,
var_spec[i].first_value);
}
return casereader_create_sequential
(NULL,
- r->value_cnt,
+ r->proto,
n_cases,
&gnm_file_casereader_class, r);
};
-/* Reads one case from READER's file into C. Returns true only
- if successful. */
-static bool
-gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_,
- struct ccase *c)
+/* Reads and returns one case from READER's file. Returns a null
+ pointer on failure. */
+static struct ccase *
+gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_)
{
+ struct ccase *c;
int ret = 0;
struct gnumeric_reader *r = r_;
if ( !r->used_first_case )
{
- *c = r->first_case;
r->used_first_case = true;
- return true;
+ return r->first_case;
}
- case_create (c, r->value_cnt);
-
- memset (case_data_rw_idx (c, 0)->s, ' ', MAX_SHORT_STRING * r->value_cnt);
+ c = case_create (r->proto);
+ case_set_missing (c);
while ((r->state == STATE_CELL || r->state == STATE_CELLS_START )
&& r->row == current_row && (ret = xmlTextReaderRead (r->xtr)))
r->col > r->stop_col))
continue;
- if ( r->col - r->start_col >= r->value_cnt)
+ if ( r->col - r->start_col >= caseproto_get_n_widths (r->proto))
continue;
if ( r->stop_row != -1 && r->row > r->stop_row)
}
- return (ret == 1);
+ if (ret == 1)
+ return c;
+ else
+ {
+ case_unref (c);
+ return NULL;
+ }
}