#include "language/command.h"
#include "language/lexer/format-parser.h"
#include "language/lexer/lexer.h"
+#include "language/lexer/token.h"
#include "language/lexer/variable-parser.h"
#include "libpspp/array.h"
#include "libpspp/assertion.h"
/* Explicit category lists. */
CCT_NUMBER,
CCT_STRING,
- CCT_RANGE,
+ CCT_NRANGE, /* Numerical range. */
+ CCT_SRANGE, /* String range. */
CCT_MISSING,
CCT_OTHERNM,
CCT_POSTCOMPUTE,
union
{
- double number; /* CCT_NUMBER. */
+ double number; /* CCT_NUMBER. */
struct substring string; /* CCT_STRING, in dictionary encoding. */
- double range[2]; /* CCT_RANGE. */
+ double nrange[2]; /* CCT_NRANGE. */
+ struct substring srange[2]; /* CCT_SRANGE. */
struct
{
switch (cat->type)
{
case CCT_NUMBER:
- case CCT_RANGE:
+ case CCT_NRANGE:
case CCT_MISSING:
case CCT_OTHERNM:
case CCT_POSTCOMPUTE:
ss_dealloc (&cat->string);
break;
+ case CCT_SRANGE:
+ ss_dealloc (&cat->srange[0]);
+ ss_dealloc (&cat->srange[1]);
+ break;
+
case CCT_SUBTOTAL:
case CCT_TOTAL:
free (cat->total_label);
}
}
+static bool
+nullable_substring_equal (const struct substring *a,
+ const struct substring *b)
+{
+ return !a->string ? !b->string : b->string && ss_equals (*a, *b);
+}
+
static bool
ctables_category_equal (const struct ctables_category *a,
const struct ctables_category *b)
case CCT_STRING:
return ss_equals (a->string, b->string);
- case CCT_RANGE:
- return a->range[0] == b->range[0] && a->range[1] == b->range[1];
+ case CCT_NRANGE:
+ return a->nrange[0] == b->nrange[0] && a->nrange[1] == b->nrange[1];
+
+ case CCT_SRANGE:
+ return (nullable_substring_equal (&a->srange[0], &b->srange[0])
+ && nullable_substring_equal (&a->srange[1], &b->srange[1]));
case CCT_MISSING:
case CCT_OTHERNM:
}
static struct ctables_category
-cct_range (double low, double high)
+cct_nrange (double low, double high)
{
return (struct ctables_category) {
- .type = CCT_RANGE,
- .range = { low, high }
+ .type = CCT_NRANGE,
+ .nrange = { low, high }
+ };
+}
+
+static struct ctables_category
+cct_srange (struct substring low, struct substring high)
+{
+ return (struct ctables_category) {
+ .type = CCT_SRANGE,
+ .srange = { low, high }
};
}
return true;
}
+static struct substring
+parse_substring (struct lexer *lexer, struct dictionary *dict)
+{
+ struct substring s = recode_substring_pool (
+ dict_get_encoding (dict), "UTF-8", lex_tokss (lexer), NULL);
+ ss_rtrim (&s, ss_cstr (" "));
+ lex_get (lexer);
+ return s;
+}
+
static bool
ctables_table_parse_explicit_category (struct lexer *lexer,
struct dictionary *dict,
return ctables_table_parse_subtotal (lexer, true, cat);
else if (lex_match_id (lexer, "LO"))
{
- if (!lex_force_match_id (lexer, "THRU") || lex_force_num (lexer))
+ if (!lex_force_match_id (lexer, "THRU"))
+ return false;
+ if (lex_is_string (lexer))
+ {
+ struct substring sr0 = { .string = NULL };
+ struct substring sr1 = parse_substring (lexer, dict);
+ *cat = cct_srange (sr0, sr1);
+ }
+ else if (lex_force_num (lexer))
+ {
+ *cat = cct_nrange (-DBL_MAX, lex_number (lexer));
+ lex_get (lexer);
+ }
+ else
return false;
- *cat = cct_range (-DBL_MAX, lex_number (lexer));
- lex_get (lexer);
}
else if (lex_is_number (lexer))
{
if (lex_match_id (lexer, "THRU"))
{
if (lex_match_id (lexer, "HI"))
- *cat = cct_range (number, DBL_MAX);
+ *cat = cct_nrange (number, DBL_MAX);
else
{
if (!lex_force_num (lexer))
return false;
- *cat = cct_range (number, lex_number (lexer));
+ *cat = cct_nrange (number, lex_number (lexer));
lex_get (lexer);
}
}
}
else if (lex_is_string (lexer))
{
- struct substring s = recode_substring_pool (
- dict_get_encoding (dict), "UTF-8", lex_tokss (lexer), NULL);
- ss_rtrim (&s, ss_cstr (" "));
-
- *cat = (struct ctables_category) { .type = CCT_STRING, .string = s };
- lex_get (lexer);
+ struct substring s = parse_substring (lexer, dict);
+ if (lex_match_id (lexer, "THRU"))
+ {
+ if (lex_match_id (lexer, "HI"))
+ {
+ struct substring sr1 = { .string = NULL };
+ *cat = cct_srange (s, sr1);
+ }
+ else
+ {
+ if (!lex_force_string (lexer))
+ return false;
+ struct substring sr1 = parse_substring (lexer, dict);
+ *cat = cct_srange (s, sr1);
+ }
+ }
+ else
+ *cat = (struct ctables_category) { .type = CCT_STRING, .string = s };
}
else if (lex_match (lexer, T_AND))
{
break;
case CTPO_CAT_RANGE:
- if (cat->type == CCT_RANGE
- && cat->range[0] == e->range[0]
- && cat->range[1] == e->range[1])
+ if (cat->type == CCT_NRANGE
+ && cat->nrange[0] == e->range[0]
+ && cat->nrange[1] == e->range[1])
best = cat;
break;
}
}
+static bool
+parse_category_string (const struct ctables_category *cat,
+ struct substring s, struct dictionary *dict,
+ enum fmt_type format, double *n)
+{
+ union value v;
+ char *error = data_in (s, dict_get_encoding (dict), format,
+ settings_get_fmt_settings (), &v, 0, NULL);
+ if (error)
+ {
+ msg_at (SE, cat->location,
+ _("Failed to parse category specification as format %s: %s."),
+ fmt_name (format), error);
+ free (error);
+ return false;
+ }
+
+ *n = v.f;
+ return true;
+}
+
+static bool
+all_strings (struct variable **vars, size_t n_vars,
+ const struct ctables_category *cat)
+{
+ for (size_t j = 0; j < n_vars; j++)
+ if (var_is_numeric (vars[j]))
+ {
+ msg_at (SE, cat->location,
+ _("This category specification may be applied only to string "
+ "variables, but this subcommand tries to apply it to "
+ "numeric variable %s."),
+ var_get_name (vars[j]));
+ return false;
+ }
+ return true;
+}
+
static bool
ctables_table_parse_categories (struct lexer *lexer, struct dictionary *dict,
struct ctables *ct, struct ctables_table *t)
break;
case CCT_NUMBER:
- case CCT_RANGE:
+ case CCT_NRANGE:
for (size_t j = 0; j < n_vars; j++)
if (var_is_alpha (vars[j]))
{
case CCT_STRING:
if (parse_strings)
{
- union value v;
- char *error = data_in (cat->string, dict_get_encoding (dict),
- common_format->type,
- settings_get_fmt_settings (),
- &v, 0, NULL);
- if (error)
- {
- msg_at (SE, cat->location,
- _("Failed to parse category specification as "
- "format %s: %s."),
- fmt_name (common_format->type), error);
- free (error);
- return false;
- }
+ double n;
+ if (!parse_category_string (cat, cat->string, dict,
+ common_format->type, &n))
+ return false;
ss_dealloc (&cat->string);
cat->type = CCT_NUMBER;
- cat->number = v.f;
+ cat->number = n;
}
- else
+ else if (!all_strings (vars, n_vars, cat))
+ return false;
+ break;
+
+ case CCT_SRANGE:
+ if (parse_strings)
{
- for (size_t j = 0; j < n_vars; j++)
- if (var_is_numeric (vars[j]))
- {
- msg_at (SE, cat->location,
- _("This category specification may be applied "
- "only to string variables, but this "
- "subcommand tries to apply it to numeric "
- "variable %s."),
- var_get_name (vars[j]));
- return false;
- }
+ double n[2];
+
+ if (!cat->srange[0].string)
+ n[0] = -DBL_MAX;
+ else if (!parse_category_string (cat, cat->srange[0], dict,
+ common_format->type, &n[0]))
+ return false;
+
+ if (!cat->srange[1].string)
+ n[1] = DBL_MAX;
+ else if (!parse_category_string (cat, cat->srange[1], dict,
+ common_format->type, &n[1]))
+ return false;
+
+ ss_dealloc (&cat->srange[0]);
+ ss_dealloc (&cat->srange[1]);
+
+ cat->type = CCT_NRANGE;
+ cat->nrange[0] = n[0];
+ cat->nrange[1] = n[1];
}
+ else if (!all_strings (vars, n_vars, cat))
+ return false;
break;
case CCT_MISSING:
{
case CCT_NUMBER:
case CCT_STRING:
- case CCT_RANGE:
+ case CCT_NRANGE:
+ case CCT_SRANGE:
case CCT_MISSING:
case CCT_OTHERNM:
cat->subtotal = subtotal;
/* Must be equal. */
continue;
- case CCT_RANGE:
+ case CCT_NRANGE:
+ case CCT_SRANGE:
case CCT_MISSING:
case CCT_OTHERNM:
{
return d;
}
+static struct substring
+rtrim_value (const union value *v, const struct variable *var)
+{
+ struct substring s = ss_buffer (CHAR_CAST (char *, v->s),
+ var_get_width (var));
+ ss_rtrim (&s, ss_cstr (" "));
+ return s;
+}
+
+static bool
+in_string_range (const union value *v, const struct variable *var,
+ const struct substring *srange)
+{
+ struct substring s = rtrim_value (v, var);
+ return ((!srange[0].string || ss_compare (s, srange[0]) >= 0)
+ && (!srange[1].string || ss_compare (s, srange[1]) <= 0));
+}
+
static const struct ctables_category *
ctables_categories_match (const struct ctables_categories *c,
const union value *v, const struct variable *var)
break;
case CCT_STRING:
- {
- struct substring s = ss_buffer (CHAR_CAST (char *, v->s),
- var_get_width (var));
- ss_rtrim (&s, ss_cstr (" "));
- printf ("%d '%.*s' ?=? '%.*s'\n",
- var_get_width (var),
- (int) cat->string.length, cat->string.string,
- (int) s.length, s.string);
- if (ss_equals (cat->string, s))
- return cat;
- }
+ if (ss_equals (cat->string, rtrim_value (v, var)))
+ return cat;
break;
- case CCT_RANGE:
- if ((cat->range[0] == -DBL_MAX || v->f >= cat->range[0])
- && (cat->range[1] == DBL_MAX || v->f <= cat->range[1]))
+ case CCT_NRANGE:
+ if ((cat->nrange[0] == -DBL_MAX || v->f >= cat->nrange[0])
+ && (cat->nrange[1] == DBL_MAX || v->f <= cat->nrange[1]))
+ return cat;
+ break;
+
+ case CCT_SRANGE:
+ if (in_string_range (v, var, cat->srange))
return cat;
break;
if (var_missing)
is_missing = true;
- printf ("ctables_cell_insert %s: ", var_get_name (var));
cats[a][i] = ctables_categories_match (
s->table->categories[var_get_dict_index (var)], value, var);
if (!cats[a][i])
pivot_table_set_caption (
pt, pivot_value_new_user_text (t->caption, SIZE_MAX));
if (t->corner)
- pivot_table_set_caption (
+ pivot_table_set_corner_text (
pt, pivot_value_new_user_text (t->corner, SIZE_MAX));
bool summary_dimension = (t->summary_axis != t->slabels_axis
}
break;
- case CCT_RANGE:
+ case CCT_NRANGE:
assert (var_is_numeric (var));
for (const struct val_lab *vl = val_labs_first (val_labs); vl;
vl = val_labs_next (val_labs, vl))
- if (vl->value.f >= c->range[0] && vl->value.f <= c->range[1])
+ if (vl->value.f >= c->nrange[0] && vl->value.f <= c->nrange[1])
+ ctables_add_occurrence (var, &vl->value, occurrences);
+ break;
+
+ case CCT_SRANGE:
+ assert (var_is_alpha (var));
+ for (const struct val_lab *vl = val_labs_first (val_labs); vl;
+ vl = val_labs_next (val_labs, vl))
+ if (in_string_range (&vl->value, var, c->srange))
ctables_add_occurrence (var, &vl->value, occurrences);
break;
return false;
}
+static void
+put_strftime (struct string *out, time_t now, const char *format)
+{
+ const struct tm *tm = localtime (&now);
+ char value[128];
+ strftime (value, sizeof value, format, tm);
+ ds_put_cstr (out, value);
+}
+
+static bool
+skip_prefix (struct substring *s, struct substring prefix)
+{
+ if (ss_starts_with (*s, prefix))
+ {
+ ss_advance (s, prefix.length);
+ return true;
+ }
+ else
+ return false;
+}
+
+static void
+put_table_expression (struct string *out, struct lexer *lexer,
+ struct dictionary *dict, int expr_start, int expr_end)
+{
+ size_t nest = 0;
+ for (int ofs = expr_start; ofs < expr_end; ofs++)
+ {
+ const struct token *t = lex_ofs_token (lexer, ofs);
+ if (t->type == T_LBRACK)
+ nest++;
+ else if (t->type == T_RBRACK && nest > 0)
+ nest--;
+ else if (nest > 0)
+ {
+ /* Nothing. */
+ }
+ else if (t->type == T_ID)
+ {
+ const struct variable *var
+ = dict_lookup_var (dict, t->string.string);
+ const char *label = var ? var_get_label (var) : NULL;
+ ds_put_cstr (out, label ? label : t->string.string);
+ }
+ else
+ {
+ if (ofs != expr_start && t->type != T_RPAREN && ds_last (out) != ' ')
+ ds_put_byte (out, ' ');
+
+ char *repr = lex_ofs_representation (lexer, ofs, ofs);
+ ds_put_cstr (out, repr);
+ free (repr);
+
+ if (ofs + 1 != expr_end && t->type != T_LPAREN)
+ ds_put_byte (out, ' ');
+ }
+ }
+}
+
+static void
+put_title_text (struct string *out, struct substring in, time_t now,
+ struct lexer *lexer, struct dictionary *dict,
+ int expr_start, int expr_end)
+{
+ for (;;)
+ {
+ size_t chunk = ss_find_byte (in, ')');
+ ds_put_substring (out, ss_head (in, chunk));
+ ss_advance (&in, chunk);
+ if (ss_is_empty (in))
+ return;
+
+ if (skip_prefix (&in, ss_cstr (")DATE")))
+ put_strftime (out, now, "%x");
+ else if (skip_prefix (&in, ss_cstr (")TIME")))
+ put_strftime (out, now, "%X");
+ else if (skip_prefix (&in, ss_cstr (")TABLE")))
+ put_table_expression (out, lexer, dict, expr_start, expr_end);
+ else
+ {
+ ds_put_byte (out, ')');
+ ss_advance (&in, 1);
+ }
+ }
+}
+
int
cmd_ctables (struct lexer *lexer, struct dataset *ds)
{
.postcomputes = HMAP_INITIALIZER (ct->postcomputes),
};
+ time_t now = time (NULL);
+
struct ctf
{
enum fmt_type type;
ct->tables[ct->n_tables++] = t;
lex_match (lexer, T_EQUALS);
+ int expr_start = lex_ofs (lexer);
if (!ctables_axis_parse (lexer, dataset_dict (ds), ct, t, PIVOT_AXIS_ROW))
goto error;
if (lex_match (lexer, T_BY))
goto error;
}
}
+ int expr_end = lex_ofs (lexer);
if (!t->axes[PIVOT_AXIS_ROW] && !t->axes[PIVOT_AXIS_COLUMN]
&& !t->axes[PIVOT_AXIS_LAYER])
{
if (!ds_is_empty (&s))
ds_put_byte (&s, ' ');
- ds_put_substring (&s, lex_tokss (lexer));
+ put_title_text (&s, lex_tokss (lexer), now,
+ lexer, dataset_dict (ds),
+ expr_start, expr_end);
lex_get (lexer);
}
free (*textp);
projects / pspp / blobdiff