X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fdata%2Fdictionary.c;h=cee678e407878651360979f77b290e614968204d;hb=37273a2e5b058a1907bc7b4b5bf666c64e0afdbb;hp=ca1a286c4d135ad20345fabe11484242bac3d9af;hpb=077a1c38bd58911cb74a08f95be3691e49b87779;p=pspp-builds.git
diff --git a/src/data/dictionary.c b/src/data/dictionary.c
index ca1a286c..cee678e4 100644
--- a/src/data/dictionary.c
+++ b/src/data/dictionary.c
@@ -1,20 +1,18 @@
-/* PSPP - computes sample statistics.
+/* PSPP - a program for statistical analysis.
Copyright (C) 1997-9, 2000, 2006, 2007 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 the Free Software Foundation; either version 2 of the
- License, or (at your option) any later version.
+ 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
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
- This program is distributed in the hope that it will be useful, but
- WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- General Public License for more details.
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
- 02110-1301, USA. */
+ along with this program. If not, see . */
#include
@@ -24,7 +22,6 @@
#include
#include "case.h"
-#include "cat-routines.h"
#include "category.h"
#include "settings.h"
#include "value-labels.h"
@@ -52,13 +49,13 @@ struct dictionary
size_t var_cnt, var_cap; /* Number of variables, capacity. */
struct hsh_table *name_tab; /* Variable index by name. */
int next_value_idx; /* Index of next `union value' to allocate. */
- struct variable **split; /* SPLIT FILE vars. */
+ const struct variable **split; /* SPLIT FILE vars. */
size_t split_cnt; /* SPLIT FILE count. */
struct variable *weight; /* WEIGHT variable. */
struct variable *filter; /* FILTER variable. */
size_t case_limit; /* Current case limit (N command). */
char *label; /* File label. */
- char *documents; /* Documents, as a string. */
+ struct string documents; /* Documents, as a string. */
struct vector **vector; /* Vectors of variables. */
size_t vector_cnt; /* Number of vectors. */
const struct dict_callbacks *callbacks; /* Callbacks on dictionary
@@ -117,7 +114,10 @@ dict_clone (const struct dictionary *s)
{
struct variable *sv = s->var[i];
struct variable *dv = dict_clone_var_assert (d, sv, var_get_name (sv));
- var_set_short_name (dv, var_get_short_name (sv));
+ size_t i;
+
+ for (i = 0; i < var_get_short_name_cnt (sv); i++)
+ var_set_short_name (dv, i, var_get_short_name (sv, i));
}
d->next_value_idx = s->next_value_idx;
@@ -159,14 +159,22 @@ dict_clear (struct dictionary *d)
while (d->var_cnt > 0 )
{
- var_clear_vardict (d->var[d->var_cnt - 1]);
- var_destroy (d->var[d->var_cnt -1]);
+ struct variable *v = d->var[d->var_cnt - 1];
+ int dict_index = var_get_dict_index (v);
+ int case_index = var_get_case_index (v);
+ int val_cnt = var_get_value_cnt (v);
+
+ var_clear_vardict (v);
+ var_destroy (v);
d->var_cnt--;
if (d->callbacks && d->callbacks->var_deleted )
- d->callbacks->var_deleted (d, d->var_cnt, d->cb_data);
+ d->callbacks->var_deleted (d,
+ dict_index, case_index, val_cnt,
+ d->cb_data);
}
+
free (d->var);
d->var = NULL;
d->var_cnt = d->var_cap = 0;
@@ -178,8 +186,7 @@ dict_clear (struct dictionary *d)
d->case_limit = 0;
free (d->label);
d->label = NULL;
- free (d->documents);
- d->documents = NULL;
+ ds_destroy (&d->documents);
dict_clear_vectors (d);
}
@@ -233,13 +240,20 @@ dict_get_var (const struct dictionary *d, size_t idx)
return d->var[idx];
}
+inline void
+dict_get_vars (const struct dictionary *d, const struct variable ***vars,
+ size_t *cnt, unsigned exclude_classes)
+{
+ dict_get_vars_mutable (d, (struct variable ***) vars, cnt, exclude_classes);
+}
+
/* Sets *VARS to an array of pointers to variables in D and *CNT
to the number of variables in *D. All variables are returned
if EXCLUDE_CLASSES is 0, or it may contain one or more of (1u
<< DC_ORDINARY), (1u << DC_SYSTEM), or (1u << DC_SCRATCH) to
exclude the corresponding type of variable. */
void
-dict_get_vars (const struct dictionary *d, struct variable ***vars,
+dict_get_vars_mutable (const struct dictionary *d, struct variable ***vars,
size_t *cnt, unsigned exclude_classes)
{
size_t count;
@@ -415,7 +429,7 @@ set_var_dict_index (struct variable *v, int dict_index)
d->callbacks->var_changed (d, dict_index, d->cb_data);
}
-/* Sets the case_index in V's vardict to DICT_INDEX. */
+/* Sets the case_index in V's vardict to CASE_INDEX. */
static void
set_var_case_index (struct variable *v, int case_index)
{
@@ -438,11 +452,12 @@ reindex_vars (struct dictionary *d, size_t from, size_t to)
/* Deletes variable V from dictionary D and frees V.
This is a very bad idea if there might be any pointers to V
- from outside D. In general, no variable in should be deleted when
- any transformations are active on the dictionary's dataset, because
- those transformations might reference the deleted variable.
- The safest time to delete a variable is just after a procedure
- has been executed, as done by MODIFY VARS.
+ from outside D. In general, no variable in the active file's
+ dictionary should be deleted when any transformations are
+ active on the dictionary's dataset, because those
+ transformations might reference the deleted variable. The
+ safest time to delete a variable is just after a procedure has
+ been executed, as done by MODIFY VARS.
Pointers to V within D are not a problem, because
dict_delete_var() knows to remove V from split variables,
@@ -451,6 +466,8 @@ void
dict_delete_var (struct dictionary *d, struct variable *v)
{
int dict_index = var_get_dict_index (v);
+ const int case_index = var_get_case_index (v);
+ const int val_cnt = var_get_value_cnt (v);
assert (dict_contains_var (d, v));
@@ -483,7 +500,7 @@ dict_delete_var (struct dictionary *d, struct variable *v)
var_destroy (v);
if (d->callbacks && d->callbacks->var_deleted )
- d->callbacks->var_deleted (d, dict_index, d->cb_data);
+ d->callbacks->var_deleted (d, dict_index, case_index, val_cnt, d->cb_data);
}
/* Deletes the COUNT variables listed in VARS from D. This is
@@ -610,7 +627,7 @@ dict_rename_var (struct dictionary *d, struct variable *v,
hsh_force_insert (d->name_tab, v);
if (get_algorithm () == ENHANCED)
- var_clear_short_name (v);
+ var_clear_short_names (v);
if ( d->callbacks && d->callbacks->var_changed )
d->callbacks->var_changed (d, var_get_dict_index (v), d->cb_data);
@@ -676,7 +693,7 @@ dict_rename_vars (struct dictionary *d,
/* Clear short names. */
if (get_algorithm () == ENHANCED)
for (i = 0; i < count; i++)
- var_clear_short_name (vars[i]);
+ var_clear_short_names (vars[i]);
pool_destroy (pool);
return true;
@@ -712,7 +729,7 @@ dict_get_case_weight (const struct dictionary *d, const struct ccase *c,
double w = case_num (c, d->weight);
if (w < 0.0 || var_is_num_missing (d->weight, w, MV_ANY))
w = 0.0;
- if ( w == 0.0 && *warn_on_invalid ) {
+ if ( w == 0.0 && warn_on_invalid != NULL && *warn_on_invalid ) {
*warn_on_invalid = false;
msg (SW, _("At least one case in the data file had a weight value "
"that was user-missing, system-missing, zero, or "
@@ -830,6 +847,33 @@ dict_compact_values (struct dictionary *d)
}
}
+
+/*
+ Reassigns case indices for D, increasing each index above START by
+ the value PADDING.
+*/
+static void
+dict_pad_values (struct dictionary *d, int start, int padding)
+{
+ size_t i;
+
+ if ( padding <= 0 )
+ return;
+
+ for (i = 0; i < d->var_cnt; ++i)
+ {
+ struct variable *v = d->var[i];
+
+ int index = var_get_case_index (v);
+
+ if ( index >= start)
+ set_var_case_index (v, index + padding);
+ }
+
+ d->next_value_idx += padding;
+}
+
+
/* Returns the number of values that would be used by a case if
dict_compact_values() were called. */
size_t
@@ -1009,7 +1053,7 @@ dict_compactor_destroy (struct dict_compactor *compactor)
dict_get_split_cnt() to determine how many SPLIT FILE vars
there are. Returns a null pointer if and only if there are no
SPLIT FILE vars. */
-struct variable *const *
+const struct variable *const *
dict_get_split_vars (const struct dictionary *d)
{
assert (d != NULL);
@@ -1051,7 +1095,7 @@ dict_set_split_vars (struct dictionary *d,
assert (cnt == 0 || split != NULL);
d->split_cnt = cnt;
- d->split = xnrealloc (d->split, cnt, sizeof *d->split);
+ d->split = cnt > 0 ? xnrealloc (d->split, cnt, sizeof *d->split) : NULL;
memcpy (d->split, split, cnt * sizeof *d->split);
if ( d->callbacks && d->callbacks->split_changed )
@@ -1089,27 +1133,73 @@ dict_set_label (struct dictionary *d, const char *label)
}
/* Returns the documents for D, or a null pointer if D has no
- documents (see cmd_document()).. */
+ documents. If the return value is nonnull, then the string
+ will be an exact multiple of DOC_LINE_LENGTH bytes in length,
+ with each segment corresponding to one line. */
const char *
dict_get_documents (const struct dictionary *d)
{
- assert (d != NULL);
-
- return d->documents;
+ return ds_is_empty (&d->documents) ? NULL : ds_cstr (&d->documents);
}
/* Sets the documents for D to DOCUMENTS, or removes D's
- documents if DOCUMENT is a null pointer. */
+ documents if DOCUMENT is a null pointer. If DOCUMENTS is
+ nonnull, then it should be an exact multiple of
+ DOC_LINE_LENGTH bytes in length, with each segment
+ corresponding to one line. */
void
dict_set_documents (struct dictionary *d, const char *documents)
{
- assert (d != NULL);
+ size_t remainder;
- free (d->documents);
- if (documents == NULL)
- d->documents = NULL;
- else
- d->documents = xstrdup (documents);
+ ds_assign_cstr (&d->documents, documents != NULL ? documents : "");
+
+ /* In case the caller didn't get it quite right, pad out the
+ final line with spaces. */
+ remainder = ds_length (&d->documents) % DOC_LINE_LENGTH;
+ if (remainder != 0)
+ ds_put_char_multiple (&d->documents, ' ', DOC_LINE_LENGTH - remainder);
+}
+
+/* Drops the documents from dictionary D. */
+void
+dict_clear_documents (struct dictionary *d)
+{
+ ds_clear (&d->documents);
+}
+
+/* Appends LINE to the documents in D. LINE will be truncated or
+ padded on the right with spaces to make it exactly
+ DOC_LINE_LENGTH bytes long. */
+void
+dict_add_document_line (struct dictionary *d, const char *line)
+{
+ if (strlen (line) > DOC_LINE_LENGTH)
+ {
+ /* Note to translators: "bytes" is correct, not characters */
+ msg (SW, _("Truncating document line to %d bytes."), DOC_LINE_LENGTH);
+ }
+ buf_copy_str_rpad (ds_put_uninit (&d->documents, DOC_LINE_LENGTH),
+ DOC_LINE_LENGTH, line);
+}
+
+/* Returns the number of document lines in dictionary D. */
+size_t
+dict_get_document_line_cnt (const struct dictionary *d)
+{
+ return ds_length (&d->documents) / DOC_LINE_LENGTH;
+}
+
+/* Copies document line number IDX from dictionary D into
+ LINE, trimming off any trailing white space. */
+void
+dict_get_document_line (const struct dictionary *d,
+ size_t idx, struct string *line)
+{
+ assert (idx < dict_get_document_line_cnt (d));
+ ds_assign_substring (line, ds_substr (&d->documents, idx * DOC_LINE_LENGTH,
+ DOC_LINE_LENGTH));
+ ds_rtrim (line, ss_cstr (CC_SPACES));
}
/* Creates in D a vector named NAME that contains the CNT
@@ -1195,137 +1285,28 @@ dict_clear_vectors (struct dictionary *d)
d->vector_cnt = 0;
}
-/* Compares two strings. */
-static int
-compare_strings (const void *a, const void *b, const void *aux UNUSED)
-{
- return strcmp (a, b);
-}
-
-/* Hashes a string. */
-static unsigned
-hash_string (const void *s, const void *aux UNUSED)
-{
- return hsh_hash_string (s);
-}
-
-
-/* Sets V's short name to BASE, followed by a suffix of the form
- _A, _B, _C, ..., _AA, _AB, etc. according to the value of
- SUFFIX_NUMBER. Truncates BASE as necessary to fit. */
-static void
-set_var_short_name_suffix (struct variable *v, const char *base,
- int suffix_number)
-{
- char suffix[SHORT_NAME_LEN + 1];
- char short_name[SHORT_NAME_LEN + 1];
- char *start, *end;
- int len, ofs;
-
- assert (v != NULL);
- assert (suffix_number >= 0);
-
- /* Set base name. */
- var_set_short_name (v, base);
-
- /* Compose suffix. */
- start = end = suffix + sizeof suffix - 1;
- *end = '\0';
- do
- {
- *--start = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"[suffix_number % 26];
- if (start <= suffix + 1)
- msg (SE, _("Variable suffix too large."));
- suffix_number /= 26;
- }
- while (suffix_number > 0);
- *--start = '_';
-
- /* Append suffix to V's short name. */
- str_copy_trunc (short_name, sizeof short_name, base);
- len = end - start;
- if (len + strlen (short_name) > SHORT_NAME_LEN)
- ofs = SHORT_NAME_LEN - len;
- else
- ofs = strlen (short_name);
- strcpy (short_name + ofs, start);
-
- /* Set name. */
- var_set_short_name (v, short_name);
-}
-
-/* Assigns a valid, unique short_name[] to each variable in D.
- Each variable whose actual name is short has highest priority
- for that short name. Otherwise, variables with an existing
- short_name[] have the next highest priority for a given short
- name; if it is already taken, then the variable is treated as
- if short_name[] had been empty. Otherwise, long names are
- truncated to form short names. If that causes conflicts,
- variables are renamed as PREFIX_A, PREFIX_B, and so on. */
+/* Called from variable.c to notify the dictionary that some property of
+ the variable has changed */
void
-dict_assign_short_names (struct dictionary *d)
+dict_var_changed (const struct variable *v)
{
- struct hsh_table *short_names;
- size_t i;
-
- /* Give variables whose names are short the corresponding short
- names, and clear short_names[] that conflict with a variable
- name. */
- for (i = 0; i < d->var_cnt; i++)
+ if ( var_has_vardict (v))
{
- struct variable *v = d->var[i];
- const char *short_name = var_get_short_name (v);
- if (strlen (var_get_name (v)) <= SHORT_NAME_LEN)
- var_set_short_name (v, var_get_name (v));
- else if (short_name != NULL && dict_lookup_var (d, short_name) != NULL)
- var_clear_short_name (v);
- }
+ const struct vardict_info *vdi = var_get_vardict (v);
+ struct dictionary *d;
- /* Each variable with an assigned short_name[] now gets it
- unless there is a conflict. */
- short_names = hsh_create (d->var_cnt, compare_strings, hash_string,
- NULL, NULL);
- for (i = 0; i < d->var_cnt; i++)
- {
- struct variable *v = d->var[i];
- const char *name = var_get_short_name (v);
- if (name != NULL && hsh_insert (short_names, (char *) name) != NULL)
- var_clear_short_name (v);
- }
+ d = vdi->dict;
- /* Now assign short names to remaining variables. */
- for (i = 0; i < d->var_cnt; i++)
- {
- struct variable *v = d->var[i];
- const char *name = var_get_short_name (v);
- if (name == NULL)
- {
- /* Form initial short_name from the variable name, then
- try _A, _B, ... _AA, _AB, etc., if needed.*/
- int trial = 0;
- do
- {
- if (trial == 0)
- var_set_short_name (v, var_get_name (v));
- else
- set_var_short_name_suffix (v, var_get_name (v), trial - 1);
-
- trial++;
- }
- while (hsh_insert (short_names, (char *) var_get_short_name (v))
- != NULL);
- }
+ if ( d->callbacks && d->callbacks->var_changed )
+ d->callbacks->var_changed (d, var_get_dict_index (v), d->cb_data);
}
-
- /* Get rid of hash table. */
- hsh_destroy (short_names);
}
-/* Called from variable.c to notify the dictionary that some property of
- the variable has changed */
+/* Called from variable.c to notify the dictionary that the variable's width
+ has changed */
void
-dict_var_changed (const struct variable *v)
+dict_var_resized (const struct variable *v, int delta)
{
if ( var_has_vardict (v))
{
@@ -1334,7 +1315,9 @@ dict_var_changed (const struct variable *v)
d = vdi->dict;
- if ( d->callbacks && d->callbacks->var_changed )
- d->callbacks->var_changed (d, var_get_dict_index (v), d->cb_data);
+ dict_pad_values (d, var_get_case_index(v) + 1, delta);
+
+ if ( d->callbacks && d->callbacks->var_resized )
+ d->callbacks->var_resized (d, var_get_dict_index (v), delta, d->cb_data);
}
}