const struct ctables_cell *example;
double d_valid; /* Dictionary weight. */
- double d_missing;
+ double d_count;
+ double d_total;
double e_valid; /* Effective weight */
- double e_missing;
+ double e_count;
+ double e_total;
};
enum ctables_summary_variant
struct ctables_domain *domains[N_CTDTS];
bool hide;
+
bool postcompute;
enum ctables_summary_variant sv;
size_t n;
size_t allocated;
+ /* The variable to which the summary specs are applied. */
struct variable *var;
+
+ /* Whether the variable to which the summary specs are applied is a scale
+ variable for the purpose of summarization.
+
+ (VALIDN and TOTALN act differently for summarizing scale and categorical
+ variables.) */
+ bool is_scale;
};
static void ctables_summary_spec_set_clone (struct ctables_summary_spec_set *,
CCT_VALUE,
CCT_LABEL,
CCT_FUNCTION,
+
+ /* For contributing to TOTALN. */
+ CCT_EXCLUDED_MISSING,
}
type;
};
/* Source location. This is null for CCT_TOTAL, CCT_VALUE, CCT_LABEL,
- CCT_FUNCTION. */
+ CCT_FUNCTION, CCT_EXCLUDED_MISSING. */
struct msg_location *location;
};
case CCT_LABEL:
case CCT_FUNCTION:
break;
+
+ case CCT_EXCLUDED_MISSING:
+ break;
}
}
&& a->sort_function == b->sort_function
&& a->sort_var == b->sort_var
&& a->percentile == b->percentile);
+
+ case CCT_EXCLUDED_MISSING:
+ return true;
}
NOT_REACHED ();
.specs = specs,
.n = src->n,
.allocated = src->n,
- .var = src->var
+ .var = src->var,
+ .is_scale = src->is_scale,
};
}
break;
case CTFA_SCALE:
+#if 0
if (!axis->scale)
{
msg_at (SE, loc,
var_name);
return false;
}
+#endif
break;
case CTFA_ALL:
case CCT_VALUE:
case CCT_LABEL:
case CCT_FUNCTION:
+ case CCT_EXCLUDED_MISSING:
break;
}
}
{
ctables_summary_spec_set_clone (&nest->specs[sv], &a->specs[sv]);
nest->specs[sv].var = a->var.var;
+ nest->specs[sv].is_scale = a->scale;
}
return (struct ctables_stack) { .nests = nest, .n = 1 };
union ctables_summary
{
/* COUNT, VALIDN, TOTALN. */
- struct
- {
- double valid;
- double missing;
- };
+ double count;
/* MINIMUM, MAXIMUM, RANGE. */
struct
case CTSF_ETOTALN:
case CTSF_VALIDN:
case CTSF_EVALIDN:
- s->missing = s->valid = 0;
+ s->count = 0;
break;
case CTSF_MAXIMUM:
ctables_summary_add (union ctables_summary *s,
const struct ctables_summary_spec *ss,
const struct variable *var, const union value *value,
+ bool is_scale, bool is_missing, bool excluded_missing,
double d_weight, double e_weight)
{
+ /* To determine whether a case is included in a given table for a particular
+ kind of summary, consider the following charts for each variable in the
+ table. Only if "yes" appears for every variable for the summary is the
+ case counted.
+
+ Categorical variables: VALIDN COUNT TOTALN
+ Valid values in included categories yes yes yes
+ Missing values in included categories --- yes yes
+ Missing values in excluded categories --- --- yes
+ Valid values in excluded categories --- --- ---
+
+ Scale variables: VALIDN COUNT TOTALN
+ Valid value yes yes yes
+ Missing value --- yes yes
+
+ Missing values include both user- and system-missing. (The system-missing
+ value is always in an excluded category.)
+ */
switch (ss->function)
{
- case CTSF_COUNT:
case CSTF_TOTALN:
- case CTSF_VALIDN:
- if (var_is_value_missing (var, value))
- s->missing += d_weight;
- else
- s->valid += d_weight;
+ case CTSF_ROWPCT_TOTALN:
+ case CTSF_COLPCT_TOTALN:
+ case CTSF_TABLEPCT_TOTALN:
+ case CTSF_SUBTABLEPCT_TOTALN:
+ case CTSF_LAYERPCT_TOTALN:
+ case CTSF_LAYERROWPCT_TOTALN:
+ case CTSF_LAYERCOLPCT_TOTALN:
+ s->count += d_weight;
break;
- case CTSF_ECOUNT:
+ case CTSF_COUNT:
case CTSF_ROWPCT_COUNT:
case CTSF_COLPCT_COUNT:
case CTSF_TABLEPCT_COUNT:
case CTSF_LAYERPCT_COUNT:
case CTSF_LAYERROWPCT_COUNT:
case CTSF_LAYERCOLPCT_COUNT:
+ if (is_scale || !excluded_missing)
+ s->count += d_weight;
+ break;
+
+ case CTSF_VALIDN:
case CTSF_ROWPCT_VALIDN:
case CTSF_COLPCT_VALIDN:
case CTSF_TABLEPCT_VALIDN:
case CTSF_LAYERPCT_VALIDN:
case CTSF_LAYERROWPCT_VALIDN:
case CTSF_LAYERCOLPCT_VALIDN:
- case CTSF_ROWPCT_TOTALN:
- case CTSF_COLPCT_TOTALN:
- case CTSF_TABLEPCT_TOTALN:
- case CTSF_SUBTABLEPCT_TOTALN:
- case CTSF_LAYERPCT_TOTALN:
- case CTSF_LAYERROWPCT_TOTALN:
- case CTSF_LAYERCOLPCT_TOTALN:
+ if (is_scale
+ ? !var_is_value_missing (var, value)
+ : !is_missing)
+ s->count += d_weight;
+ break;
+
case CTSF_MISSING:
- case CTSF_ETOTALN:
+ if (is_missing)
+ s->count += d_weight;
+ break;
+
+ case CTSF_ECOUNT:
+ if (is_scale || !excluded_missing)
+ s->count += e_weight;
+ break;
+
case CTSF_EVALIDN:
- if (var_is_value_missing (var, value))
- s->missing += e_weight;
- else
- s->valid += e_weight;
+ if (is_scale
+ ? !var_is_value_missing (var, value)
+ : !is_missing)
+ s->count += e_weight;
+ break;
+
+ case CTSF_ETOTALN:
+ s->count += e_weight;
break;
case CTSF_MAXIMUM:
{
case CTSF_COUNT:
case CTSF_ECOUNT:
- return s->valid;
+ return s->count;
case CTSF_ROWPCT_COUNT:
case CTSF_COLPCT_COUNT:
case CTSF_LAYERCOLPCT_COUNT:
{
enum ctables_domain_type d = ctables_function_domain (ss->function);
- return (cell->domains[d]->e_valid
- ? s->valid / cell->domains[d]->e_valid * 100
+ return (cell->domains[d]->e_count
+ ? s->count / cell->domains[d]->e_count * 100
: SYSMIS);
}
case CTSF_LAYERPCT_VALIDN:
case CTSF_LAYERROWPCT_VALIDN:
case CTSF_LAYERCOLPCT_VALIDN:
+ {
+ enum ctables_domain_type d = ctables_function_domain (ss->function);
+ return (cell->domains[d]->e_valid
+ ? s->count / cell->domains[d]->e_valid * 100
+ : SYSMIS);
+ }
+
case CTSF_ROWPCT_TOTALN:
case CTSF_COLPCT_TOTALN:
case CTSF_TABLEPCT_TOTALN:
case CTSF_LAYERPCT_TOTALN:
case CTSF_LAYERROWPCT_TOTALN:
case CTSF_LAYERCOLPCT_TOTALN:
- NOT_REACHED ();
+ {
+ enum ctables_domain_type d = ctables_function_domain (ss->function);
+ return (cell->domains[d]->e_total
+ ? s->count / cell->domains[d]->e_total * 100
+ : SYSMIS);
+ }
case CTSF_MISSING:
- return s->missing;
+ return s->count;
case CSTF_TOTALN:
case CTSF_ETOTALN:
- return s->valid + s->missing;
+ return s->count;
case CTSF_VALIDN:
+ return s->count;
+
case CTSF_EVALIDN:
- return s->valid;
+ return s->count;
case CTSF_MAXIMUM:
return s->max;
case CCT_SUBTOTAL:
case CCT_TOTAL:
case CCT_POSTCOMPUTE:
+ case CCT_EXCLUDED_MISSING:
/* Must be equal. */
continue;
ctables_categories_match (const struct ctables_categories *c,
const union value *v, const struct variable *var)
{
+ if (var_is_numeric (var) && v->f == SYSMIS)
+ return NULL;
+
const struct ctables_category *othernm = NULL;
for (size_t i = c->n_cats; i-- > 0; )
{
case CCT_FUNCTION:
return (cat->include_missing || !var_is_value_missing (var, v) ? cat
: NULL);
+
+ case CCT_EXCLUDED_MISSING:
+ break;
}
}
for (size_t i = 0; i < nest->n; i++)
{
const struct ctables_category *cat = cats[a][i];
+ const struct variable *var = nest->vars[i];
+ const union value *value = case_data (c, var);
if (i != nest->scale_idx)
{
const struct ctables_category *subtotal = cat->subtotal;
}
cell->axes[a].cvs[i].category = cat;
- value_clone (&cell->axes[a].cvs[i].value, case_data (c, nest->vars[i]),
- var_get_width (nest->vars[i]));
+ value_clone (&cell->axes[a].cvs[i].value, value, var_get_width (var));
}
}
static void
ctables_cell_add__ (struct ctables_section *s, const struct ccase *c,
const struct ctables_category *cats[PIVOT_N_AXES][10],
+ bool is_missing, bool excluded_missing,
double d_weight, double e_weight)
{
struct ctables_cell *cell = ctables_cell_insert__ (s, c, cats);
const struct ctables_summary_spec_set *specs = &ss->specs[cell->sv];
for (size_t i = 0; i < specs->n; i++)
- ctables_summary_add (&cell->summaries[i], &specs->specs[i], specs->var,
- case_data (c, specs->var), d_weight, e_weight);
+ ctables_summary_add (&cell->summaries[i], &specs->specs[i],
+ specs->var, case_data (c, specs->var), specs->is_scale,
+ is_missing, excluded_missing, d_weight, e_weight);
if (cell->contributes_to_domains)
{
for (enum ctables_domain_type dt = 0; dt < N_CTDTS; dt++)
{
- cell->domains[dt]->d_valid += d_weight;
- cell->domains[dt]->e_valid += e_weight;
+ struct ctables_domain *d = cell->domains[dt];
+ d->d_total += d_weight;
+ d->e_total += e_weight;
+ if (!excluded_missing)
+ {
+ d->d_count += d_weight;
+ d->e_count += e_weight;
+ }
+ if (!is_missing)
+ {
+ d->d_valid += d_weight;
+ d->e_valid += e_weight;
+ }
}
}
}
static void
recurse_totals (struct ctables_section *s, const struct ccase *c,
const struct ctables_category *cats[PIVOT_N_AXES][10],
+ bool is_missing, bool excluded_missing,
double d_weight, double e_weight,
enum pivot_axis_type start_axis, size_t start_nest)
{
{
const struct ctables_category *save = cats[a][i];
cats[a][i] = total;
- ctables_cell_add__ (s, c, cats, d_weight, e_weight);
- recurse_totals (s, c, cats, d_weight, e_weight, a, i + 1);
+ ctables_cell_add__ (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight);
+ recurse_totals (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight, a, i + 1);
cats[a][i] = save;
}
}
static void
recurse_subtotals (struct ctables_section *s, const struct ccase *c,
const struct ctables_category *cats[PIVOT_N_AXES][10],
+ bool is_missing, bool excluded_missing,
double d_weight, double e_weight,
enum pivot_axis_type start_axis, size_t start_nest)
{
if (save->subtotal)
{
cats[a][i] = save->subtotal;
- ctables_cell_add__ (s, c, cats, d_weight, e_weight);
- recurse_subtotals (s, c, cats, d_weight, e_weight, a, i + 1);
+ ctables_cell_add__ (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight);
+ recurse_subtotals (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight, a, i + 1);
cats[a][i] = save;
}
}
double d_weight, double e_weight)
{
const struct ctables_category *cats[PIVOT_N_AXES][10]; /* XXX */
+
+ /* Does at least one categorical variable have a missing value in an included
+ or excluded category? */
+ bool is_missing = false;
+
+ /* Does at least one categorical variable have a missing value in an excluded
+ category? */
+ bool excluded_missing = false;
+
for (enum pivot_axis_type a = 0; a < PIVOT_N_AXES; a++)
{
const struct ctables_nest *nest = s->nests[a];
const struct variable *var = nest->vars[i];
const union value *value = case_data (c, var);
- if (var_is_numeric (var) && value->f == SYSMIS)
- return;
+ bool var_missing = var_is_value_missing (var, value) != 0;
+ if (var_missing)
+ is_missing = true;
cats[a][i] = ctables_categories_match (
s->table->categories[var_get_dict_index (var)], value, var);
if (!cats[a][i])
- return;
+ {
+ if (!var_missing)
+ return;
+
+ static const struct ctables_category cct_excluded_missing = {
+ .type = CCT_EXCLUDED_MISSING,
+ .hide = true,
+ };
+ cats[a][i] = &cct_excluded_missing;
+ excluded_missing = true;
+ }
}
}
- for (enum pivot_axis_type a = 0; a < PIVOT_N_AXES; a++)
- {
- const struct ctables_nest *nest = s->nests[a];
- for (size_t i = 0; i < nest->n; i++)
- if (i != nest->scale_idx)
- {
- const struct variable *var = nest->vars[i];
- const union value *value = case_data (c, var);
- ctables_add_occurrence (var, value, &s->occurrences[a][i]);
- }
- }
+ if (!excluded_missing)
+ for (enum pivot_axis_type a = 0; a < PIVOT_N_AXES; a++)
+ {
+ const struct ctables_nest *nest = s->nests[a];
+ for (size_t i = 0; i < nest->n; i++)
+ if (i != nest->scale_idx)
+ {
+ const struct variable *var = nest->vars[i];
+ const union value *value = case_data (c, var);
+ ctables_add_occurrence (var, value, &s->occurrences[a][i]);
+ }
+ }
- ctables_cell_add__ (s, c, cats, d_weight, e_weight);
+ ctables_cell_add__ (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight);
- recurse_totals (s, c, cats, d_weight, e_weight, 0, 0);
- recurse_subtotals (s, c, cats, d_weight, e_weight, 0, 0);
+ //if (!excluded_missing)
+ {
+ recurse_totals (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight, 0, 0);
+ recurse_subtotals (s, c, cats, is_missing, excluded_missing,
+ d_weight, e_weight, 0, 0);
+ }
}
struct merge_item
specs->n = 1;
enum ctables_summary_function function
- = specs->var ? CTSF_MEAN : CTSF_COUNT;
+ = specs->is_scale ? CTSF_MEAN : CTSF_COUNT;
struct ctables_var var = { .is_mrset = false, .var = specs->var };
*specs->specs = (struct ctables_summary_spec) {
if (c->include_missing || !var_is_value_missing (var, &vl->value))
ctables_add_occurrence (var, &vl->value, occurrences);
break;
+
+ case CCT_EXCLUDED_MISSING:
+ break;
}
}
}
projects / pspp / blobdiff