numeric user-missing values like system-missing values and string
user-missing values as all spaces.
-By default, all the variables in the active dataset dictionary are saved
-to the system file, but @subcmd{DROP} or @subcmd{KEEP} can select a subset of variable
-to save. The @subcmd{RENAME} subcommand can also be used to change the names
-under which variables are saved. @subcmd{UNSELECTED} determines whether cases
-filtered out by the @cmd{FILTER} command are written to the output file.
-These subcommands have the same syntax and meaning as on the
-@cmd{SAVE} command (@pxref{SAVE}).
+By default, all the variables in the active dataset dictionary are
+saved to the system file, but @subcmd{DROP} or @subcmd{KEEP} can
+select a subset of variable to save. The @subcmd{RENAME} subcommand
+can also be used to change the names under which variables are saved;
+because they are used only in the output, these names do not have to
+conform to the usual PSPP variable naming rules. @subcmd{UNSELECTED}
+determines whether cases filtered out by the @cmd{FILTER} command are
+written to the output file. These subcommands have the same syntax
+and meaning as on the @cmd{SAVE} command (@pxref{SAVE}).
Each supported file type has additional subcommands, explained in
separate sections below.
struct mrset **mrsets; /* Multiple response sets. */
size_t n_mrsets; /* Number of multiple response sets. */
+ /* Whether variable names must be valid identifiers. Normally, this is
+ true, but sometimes a dictionary is prepared for external use
+ (e.g. output to a CSV file) where names don't have to be valid. */
+ bool names_must_be_ids;
+
char *encoding; /* Character encoding of string data */
const struct dict_callbacks *callbacks; /* Callbacks on dictionary
dict_id_is_valid (const struct dictionary *dict, const char *id,
bool issue_error)
{
- return id_is_valid (id, dict->encoding, issue_error);
+ return (!dict->names_must_be_ids
+ || id_is_valid (id, dict->encoding, issue_error));
}
void
struct dictionary *d = xzalloc (sizeof *d);
d->encoding = xstrdup (encoding);
+ d->names_must_be_ids = true;
hmap_init (&d->name_map);
attrset_init (&d->attributes);
size_t i;
d = dict_create (s->encoding);
+ dict_set_names_must_be_ids (d, dict_get_names_must_be_ids (s));
for (i = 0; i < s->var_cnt; i++)
{
return make_numeric_name (dict, num_start);
}
+/* Returns whether variable names must be valid identifiers. Normally, this is
+ true, but sometimes a dictionary is prepared for external use (e.g. output
+ to a CSV file) where names don't have to be valid. */
+bool
+dict_get_names_must_be_ids (const struct dictionary *d)
+{
+ return d->names_must_be_ids;
+}
+
+/* Sets whether variable names must be valid identifiers. Normally, this is
+ true, but sometimes a dictionary is prepared for external use (e.g. output
+ to a CSV file) where names don't have to be valid.
+
+ Changing this setting from false to true doesn't make the dictionary check
+ all the existing variable names, so it can cause an invariant violation. */
+void
+dict_set_names_must_be_ids (struct dictionary *d, bool names_must_be_ids)
+{
+ d->names_must_be_ids = names_must_be_ids;
+}
+
/* Returns the weighting variable in dictionary D, or a null
pointer if the dictionary is unweighted. */
struct variable *
char *dict_make_unique_var_name (const struct dictionary *, const char *hint,
unsigned long int *num_start);
+bool dict_get_names_must_be_ids (const struct dictionary *);
+void dict_set_names_must_be_ids (struct dictionary *, bool);
+
/* Weight variable. */
double dict_get_case_weight (const struct dictionary *,
const struct ccase *, bool *);
var_set_name_quiet (struct variable *v, const char *name)
{
assert (!var_has_vardict (v));
- assert (id_is_plausible (name, false));
free (v->name);
v->name = xstrdup (name);
{
struct variable *ov = var_clone (var);
- assert (short_name == NULL || id_is_plausible (short_name, false));
-
/* Clear old short name numbered IDX, if any. */
if (idx < var->short_name_cnt)
{
type = 0;
dict = dict_clone (dataset_dict (ds));
+ dict_set_names_must_be_ids (dict, false);
stage = NULL;
map = NULL;
lex_match (lexer, T_EQUALS);
if (lex_token (lexer) != T_LPAREN)
{
- struct variable *v;
-
- v = parse_variable (lexer, dict);
+ struct variable *v = parse_variable (lexer, dict);
if (v == NULL)
return 0;
- if (!lex_force_match (lexer, T_EQUALS)
- || !lex_force_id (lexer)
- || !dict_id_is_valid (dict, lex_tokcstr (lexer), true))
- return 0;
- if (dict_lookup_var (dict, lex_tokcstr (lexer)) != NULL)
+ if (!lex_force_match (lexer, T_EQUALS))
+ return 0;
+
+ char *new_name = parse_DATA_LIST_var (lexer, dict);
+ if (dict_lookup_var (dict, new_name) != NULL)
{
msg (SE, _("Cannot rename %s as %s because there already exists "
"a variable named %s. To rename variables with "
"overlapping names, use a single RENAME subcommand "
"such as `/RENAME (A=B)(B=C)(C=A)', or equivalently, "
"`/RENAME (A B C=B C A)'."),
- var_get_name (v), lex_tokcstr (lexer), lex_tokcstr (lexer));
+ var_get_name (v), new_name, new_name);
+ free (new_name);
return 0;
}
- dict_rename_var (dict, v, lex_tokcstr (lexer));
- lex_get (lexer);
+ dict_rename_var (dict, v, new_name);
+ free (new_name);
return 1;
}
#include "gettext.h"
#define _(msgid) gettext (msgid)
-static struct variable * var_set_get_var (const struct var_set *, size_t );
-
+static struct variable *var_set_get_var (const struct var_set *, size_t);
static struct variable *var_set_lookup_var (const struct var_set *,
const char *);
-
static bool var_set_lookup_var_idx (const struct var_set *, const char *,
size_t *);
+static bool var_set_get_names_must_be_ids (const struct var_set *);
+static bool
+is_name_token (const struct lexer *lexer, bool names_must_be_ids)
+{
+ return (lex_token (lexer) == T_ID
+ || (!names_must_be_ids && lex_token (lexer) == T_STRING));
+}
+static bool
+is_vs_name_token (const struct lexer *lexer, const struct var_set *vs)
+{
+ return is_name_token (lexer, var_set_get_names_must_be_ids (vs));
+}
+
+static bool
+is_dict_name_token (const struct lexer *lexer, const struct dictionary *d)
+{
+ return is_name_token (lexer, dict_get_names_must_be_ids (d));
+}
/* Parses a name as a variable within VS. Sets *IDX to the
variable's index and returns true if successful. On failure
emits an error message and returns false. */
static bool
parse_vs_variable_idx (struct lexer *lexer, const struct var_set *vs,
- size_t *idx)
+ size_t *idx)
{
assert (idx != NULL);
- if (lex_token (lexer) != T_ID)
+ if (!is_vs_name_token (lexer, vs))
{
lex_error (lexer, _("expecting variable name"));
return false;
lex_match (lexer, T_COMMA);
}
while (lex_token (lexer) == T_ALL
- || (lex_token (lexer) == T_ID && var_set_lookup_var (vs, lex_tokcstr (lexer)) != NULL));
+ || (is_vs_name_token (lexer, vs)
+ && var_set_lookup_var (vs, lex_tokcstr (lexer)) != NULL));
if (*nv == 0)
goto fail;
return 0;
}
+char *
+parse_DATA_LIST_var (struct lexer *lexer, const struct dictionary *d)
+{
+ if (!is_dict_name_token (lexer, d))
+ {
+ lex_error (lexer, "expecting variable name");
+ return NULL;
+ }
+ if (!dict_id_is_valid (d, lex_tokcstr (lexer), true))
+ return NULL;
+
+ char *name = xstrdup (lex_tokcstr (lexer));
+ lex_get (lexer);
+ return name;
+}
+
/* Attempts to break UTF-8 encoded NAME into a root (whose contents are
arbitrary except that it does not end in a digit) followed by an integer
numeric suffix. On success, stores the value of the suffix into *NUMBERP,
do
{
- if (lex_token (lexer) != T_ID
- || !dict_id_is_valid (dict, lex_tokcstr (lexer), true))
- {
- lex_error (lexer, "expecting variable name");
- goto exit;
- }
- if (dict_class_from_id (lex_tokcstr (lexer)) == DC_SCRATCH
- && (pv_opts & PV_NO_SCRATCH))
+ name1 = parse_DATA_LIST_var (lexer, dict);
+ if (!name1)
+ goto exit;
+ if (dict_class_from_id (name1) == DC_SCRATCH && pv_opts & PV_NO_SCRATCH)
{
msg (SE, _("Scratch variables not allowed here."));
goto exit;
}
- name1 = xstrdup (lex_tokcstr (lexer));
- lex_get (lexer);
- if (lex_token (lexer) == T_TO)
+ if (lex_match (lexer, T_TO))
{
- char *name2 = NULL;
unsigned long int num1, num2;
int n_digits1, n_digits2;
int root_len1, root_len2;
unsigned long int number;
- lex_get (lexer);
- if (lex_token (lexer) != T_ID
- || !dict_id_is_valid (dict, lex_tokcstr (lexer), true))
- {
- lex_error (lexer, "expecting variable name");
- goto exit;
- }
- name2 = xstrdup (lex_tokcstr (lexer));
- lex_get (lexer);
+ char *name2 = parse_DATA_LIST_var (lexer, dict);
root_len1 = extract_numeric_suffix (name1, &num1, &n_digits1);
if (root_len1 == 0)
*names = NULL;
*nnames = 0;
}
- while (lex_token (lexer) == T_ID || lex_token (lexer) == T_ALL)
+ while (is_dict_name_token (lexer, dict) || lex_token (lexer) == T_ALL)
{
if (lex_token (lexer) == T_ALL || dict_lookup_var (dict, lex_tokcstr (lexer)) != NULL)
{
/* A set of variables. */
struct var_set
{
+ bool names_must_be_ids;
size_t (*get_cnt) (const struct var_set *);
struct variable *(*get_var) (const struct var_set *, size_t idx);
bool (*lookup_var_idx) (const struct var_set *, const char *, size_t *);
if (vs != NULL)
vs->destroy (vs);
}
+
+static bool
+var_set_get_names_must_be_ids (const struct var_set *vs)
+{
+ return vs->names_must_be_ids;
+}
\f
/* Returns the number of variables in VS. */
static size_t
var_set_create_from_dict (const struct dictionary *d)
{
struct var_set *vs = xmalloc (sizeof *vs);
+ vs->names_must_be_ids = dict_get_names_must_be_ids (d);
vs->get_cnt = dict_var_set_get_cnt;
vs->get_var = dict_var_set_get_var;
vs->lookup_var_idx = dict_var_set_lookup_var_idx;
size_t i;
vs = xmalloc (sizeof *vs);
+ vs->names_must_be_ids = true;
vs->get_cnt = array_var_set_get_cnt;
vs->get_var = array_var_set_get_var;
vs->lookup_var_idx = array_var_set_lookup_var_idx;
struct variable ***, size_t *, int opts);
bool parse_var_set_vars (struct lexer *, const struct var_set *, struct variable ***, size_t *,
int opts);
+
+char *parse_DATA_LIST_var (struct lexer *, const struct dictionary *);
bool parse_DATA_LIST_vars (struct lexer *, const struct dictionary *,
char ***names, size_t *cnt, int opts);
bool parse_DATA_LIST_vars_pool (struct lexer *, const struct dictionary *,
AT_SETUP([CSV output -- KEEP, RENAME])
PREPARE_SAVE_TRANSLATE_CSV(
- [/FIELDNAMES /KEEP=time string /RENAME string=name /UNSELECTED=DELETE])
+ [/FIELDNAMES /KEEP=time string /RENAME string='long name with spaces' /UNSELECTED=DELETE])
AT_CHECK([cat data.csv], [0], [dnl
-time,name
+time,long name with spaces
-05:17:00, xxx
12:00:00,xyzzy
])
projects / pspp / commitdiff