1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2017-2018 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
17 #ifndef OUTPUT_PIVOT_TABLE_H
18 #define OUTPUT_PIVOT_TABLE_H 1
22 #include "data/format.h"
23 #include "data/settings.h"
24 #include "libpspp/compiler.h"
25 #include "libpspp/hmap.h"
26 #include "output/table.h"
34 Pivot tables are PSPP's primary form of output. They are analogous to the
35 pivot tables you might be familiar with from spreadsheets and databases.
36 See https://en.wikipedia.org/wiki/Pivot_table for a brief introduction to
37 the overall concept of a pivot table.
39 In PSPP, the most important internal pieces of a pivot table are:
41 - Title. Every pivot table has a title that is displayed above it. It also
42 has an optional caption (displayed below it) and corner text (displayed in
43 the upper left corner).
45 - Dimensions. A dimension consists of zero or more categories. A category
46 has a label, such as "df" or "Asymp. Sig." or 123 or a variable name. The
47 categories are the leaves of a tree whose non-leaf nodes form groups of
48 categories. The tree always has a root group whose label is the name of
51 - Axes. A table has three axes: column, row, and layer. Each dimension is
52 assigned to an axis, and each axis has zero or more dimensions. When an
53 axis has more than one dimension, they are ordered from innermost to
56 - Data. A table's data consists of zero or more cells. Each cell maps from
57 a category for each dimension to a value, which is commonly a number but
58 could also be a variable name or an arbitrary text string.
60 Creating a pivot table usually consists of the following steps:
62 1. Create the table with pivot_table_create(), passing in the title.
63 It's commonly useful to set up a few options at this point:
65 - If empty rows or columns should not be displayed, set ->omit_empty to
68 - Set the format to use for "count" values with
69 pivot_table_set_weight_var() or pivot_table_set_weight_format().
71 2. Create each dimension with pivot_dimension_create() and populate it with
72 categories and, possibly, with groups that contain the categories. This
73 call also assigns the dimension to an axis.
75 In simple cases, only a call to pivot_dimension_create() is needed.
76 Other functions such as pivot_category_create_group() can be used for
77 hierarchies of categories.
79 Sometimes it's easier to create categories in tandem with inserting data,
80 for example by adding a category for a variable just before inserting the
81 first cell for that variable. In that case, creating categories and
82 inserting data can be interleaved.
84 3. Insert data. For each cell, supply the category indexes, which are
85 assigned starting from 0 in the order in which the categories were
86 created in step 2, and the value to go in the cell. If the table has a
87 small, fixed number of dimensions, functions like, e.g.
88 pivot_table_put3() for 3 dimensions, can be used. The general function
89 pivot_table_put() works for other cases.
91 4. Output the table for user consumption. Use pivot_table_submit(). */
93 /* Pivot table display styling. */
95 /* Areas of a pivot table for styling purposes. */
100 PIVOT_AREA_FOOTER, /* Footnotes. */
101 PIVOT_AREA_CORNER, /* Top-left corner. */
102 PIVOT_AREA_COLUMN_LABELS,
103 PIVOT_AREA_ROW_LABELS,
105 PIVOT_AREA_LAYERS, /* Layer indication. */
109 const char *pivot_area_to_string (enum pivot_area);
110 const struct area_style *pivot_area_get_default_style (enum pivot_area);
112 /* Table borders for styling purposes. */
118 PIVOT_BORDER_OUTER_LEFT,
119 PIVOT_BORDER_OUTER_TOP,
120 PIVOT_BORDER_OUTER_RIGHT,
121 PIVOT_BORDER_OUTER_BOTTOM,
124 PIVOT_BORDER_INNER_LEFT,
125 PIVOT_BORDER_INNER_TOP,
126 PIVOT_BORDER_INNER_RIGHT,
127 PIVOT_BORDER_INNER_BOTTOM,
130 PIVOT_BORDER_DATA_LEFT,
131 PIVOT_BORDER_DATA_TOP,
134 PIVOT_BORDER_DIM_ROW_HORZ,
135 PIVOT_BORDER_DIM_ROW_VERT,
136 PIVOT_BORDER_DIM_COL_HORZ,
137 PIVOT_BORDER_DIM_COL_VERT,
140 PIVOT_BORDER_CAT_ROW_HORZ,
141 PIVOT_BORDER_CAT_ROW_VERT,
142 PIVOT_BORDER_CAT_COL_HORZ,
143 PIVOT_BORDER_CAT_COL_VERT,
148 const char *pivot_border_to_string (enum pivot_border);
149 void pivot_border_get_default_style (enum pivot_border,
150 struct table_border_style *);
152 /* Sizing for rows or columns of a rendered table. The comments below talk
153 about columns and their widths but they apply equally to rows and their
155 struct pivot_table_sizing
157 /* Minimum and maximum column width, in 1/96" units. */
160 /* Specific column widths, in 1/96" units. */
164 /* Specific page breaks: 0-based columns after which a page break must
165 occur, e.g. a value of 1 requests a break after the second column. */
169 /* Keeps: columns to keep together on a page if possible. */
170 struct pivot_keep *keeps;
174 void pivot_table_sizing_uninit (struct pivot_table_sizing *);
176 /* A set of columns to keep together on a page if possible, e.g. ofs=1, n=10
177 requests keeping together the 2nd through 11th columns. */
180 size_t ofs; /* 0-based first column. */
181 size_t n; /* Number of columns. */
195 const char *pivot_axis_type_to_string (enum pivot_axis_type);
197 /* An axis within a pivot table. */
200 /* dimensions[0] is the innermost dimension,
201 dimensions[1] is the next outer dimension,
203 dimensions[n_dimensions - 1] is the outermost dimension. */
204 struct pivot_dimension **dimensions;
207 /* The number of rows or columns along the axis,
208 that is, the product of dimension[*]->n_leaves.
209 It is 0 if any dimension has 0 leaves. */
212 /* Sum of dimensions[*]->label_depth. */
216 /* Successively assigns to INDEXES (which should be a "size_t *") each of the
217 combinations of the categories in AXIS's dimensions, in lexicographic order
218 with the innermost dimension iterating most quickly.
220 The value assigned to INDEXES is dynamically allocated. If the client
221 breaks out of the loop prematurely, it needs to free it with free(). */
222 #define PIVOT_AXIS_FOR_EACH(INDEXES, AXIS) \
223 for ((INDEXES) = NULL; \
224 ((INDEXES) = pivot_axis_iterator_next (INDEXES, AXIS)) != NULL;)
225 size_t *pivot_axis_iterator_next (size_t *indexes, const struct pivot_axis *);
229 A pivot_dimension identifies the categories associated with a single
230 dimension within a multidimensional pivot table.
232 A dimension contains a collection of categories, which are the leaves in a
235 (A dimension or a group can contain zero categories, but this is unusual.
236 If a dimension contains no categories, then its table cannot contain any
239 struct pivot_dimension
241 /* table->axes[axis_type]->dimensions[level] == dimension. */
242 struct pivot_table *table;
243 enum pivot_axis_type axis_type;
244 size_t level; /* 0 for innermost dimension within axis. */
246 /* table->dimensions[top_index] == dimension. */
249 /* Hierarchy of categories within the dimension. The groups and categories
250 are sorted in the order that should be used for display. This might be
251 different from the original order produced for output if the user
254 The root must always be a group, although it is allowed to have no
256 struct pivot_category *root;
258 /* All of the leaves reachable via the root.
260 The indexing for presentation_leaves is presentation order, thus
261 presentation_leaves[i]->presentation_index == i. This order is the same
262 as would be produced by an in-order traversal of the groups. It is the
263 order into which the user reordered or sorted the categories.
265 The indexing for data_leaves is that used for idx[] in struct
266 pivot_cell, thus data_leaves[i]->data_index == i. This might differ
267 from what an in-order traversal of 'root' would yield, if the user
268 reordered categories. */
269 struct pivot_category **data_leaves;
270 struct pivot_category **presentation_leaves;
271 size_t n_leaves, allocated_leaves;
274 bool hide_all_labels;
276 /* Number of rows or columns needed to express the labels. */
280 struct pivot_dimension *pivot_dimension_create (
281 struct pivot_table *, enum pivot_axis_type, const char *name, ...)
283 #define pivot_dimension_create(...) \
284 pivot_dimension_create(__VA_ARGS__, NULL_SENTINEL)
285 struct pivot_dimension *pivot_dimension_create__ (struct pivot_table *,
286 enum pivot_axis_type,
287 struct pivot_value *name);
289 void pivot_dimension_destroy (struct pivot_dimension *);
291 void pivot_dimension_dump (const struct pivot_dimension *, int indentation);
293 /* A pivot_category is a leaf (a category) or a group:
295 - For a leaf, neither index is SIZE_MAX.
297 - For a group, both indexes are SIZE_MAX.
299 Do not use 'subs' or 'n_subs' to determine whether a category is a group,
300 because a group may (pathologically) have no leaves. */
301 struct pivot_category
303 struct pivot_value *name;
304 struct pivot_category *parent;
305 struct pivot_dimension *dimension;
306 size_t label_depth, extra_depth;
310 If show_label is true, then the group itself has a row (or a column)
311 giving the group's name. Otherwise, the group's own name is not
313 struct pivot_category **subs; /* Child categories or groups. */
314 size_t n_subs, allocated_subs;
315 bool show_label; /* Display a label for the group itself? */
316 bool show_label_in_corner;
319 struct fmt_spec format;
320 size_t group_index; /* In ->parent->subs[]. */
321 size_t data_index; /* In ->dimension->data_leaves[]. */
322 size_t presentation_index; /* In ->dimension->presentation_leaves[]. */
326 pivot_category_is_group (const struct pivot_category *category)
328 return category->data_index == SIZE_MAX;
332 pivot_category_is_leaf (const struct pivot_category *category)
334 return !pivot_category_is_group (category);
337 /* Creating leaf categories. */
338 int pivot_category_create_leaves (struct pivot_category *parent, ...)
340 #define pivot_category_create_leaves(...) \
341 pivot_category_create_leaves(__VA_ARGS__, NULL_SENTINEL)
343 int pivot_category_create_leaf (
344 struct pivot_category *parent, struct pivot_value *name);
345 int pivot_category_create_leaf_rc (
346 struct pivot_category *parent, struct pivot_value *name, const char *rc);
348 /* Creating category groups. */
349 struct pivot_category *pivot_category_create_group (
350 struct pivot_category *parent, const char *name, ...) SENTINEL (0);
351 #define pivot_category_create_group(...) \
352 pivot_category_create_group(__VA_ARGS__, NULL_SENTINEL)
353 struct pivot_category *pivot_category_create_group__ (
354 struct pivot_category *parent, struct pivot_value *name);
356 void pivot_category_destroy (struct pivot_category *);
358 /* Pivot result classes.
360 These are used to mark leaf categories as having particular types of data,
361 to set their numeric formats. The formats that actually get used for these
362 classes are in the result_classes[] global array in pivot-table.c, except
363 that PIVOT_RC_OTHER comes from settings_get_format() and PIVOT_RC_COUNT
364 should come from the weight variable in the dataset's dictionary. */
365 #define PIVOT_RC_OTHER ("RC_OTHER")
366 #define PIVOT_RC_INTEGER ("RC_INTEGER")
367 #define PIVOT_RC_CORRELATION ("RC_CORRELATIONS")
368 #define PIVOT_RC_SIGNIFICANCE ("RC_SIGNIFICANCE")
369 #define PIVOT_RC_PERCENT ("RC_PERCENT")
370 #define PIVOT_RC_RESIDUAL ("RC_RESIDUAL")
371 #define PIVOT_RC_COUNT ("RC_COUNT")
373 bool pivot_result_class_change (const char *, const struct fmt_spec *);
375 /* A pivot table. See the top of this file for more information. */
378 /* Reference count. A pivot_table may be shared between multiple owners,
379 indicated by a reference count greater than 1. When this is the case,
380 the pivot_table must not be modified. */
383 /* Display settings. */
384 bool rotate_inner_column_labels;
385 bool rotate_outer_row_labels;
386 bool row_labels_in_corner;
387 bool show_grid_lines;
389 bool omit_empty; /* Omit empty rows and columns? */
390 size_t *current_layer; /* axis[PIVOT_AXIS_LAYER].n_dimensions elements. */
392 enum settings_value_show show_values;
393 enum settings_value_show show_variables;
394 struct fmt_spec weight_format;
396 /* Footnote display settings. */
397 bool show_numeric_markers;
398 bool footnote_marker_superscripts;
400 /* Column and row sizing and page breaks.
401 sizing[TABLE_HORZ] is for columns, sizing[TABLE_VERT] is for rows. */
402 struct pivot_table_sizing sizing[TABLE_N_AXES];
404 /* Print settings. */
405 bool print_all_layers;
406 bool paginate_layers;
407 bool shrink_to_fit[TABLE_N_AXES];
408 bool top_continuation, bottom_continuation;
410 size_t n_orphan_lines;
412 /* Format settings. */
414 char decimal; /* Usually ',' or '.'. */
415 char grouping; /* Usually '.' or ','. */
416 char *ccs[5]; /* Custom currency. */
419 /* Command information. */
420 char *command_local; /* May be NULL. */
421 char *command_c; /* May be NULL. */
422 char *language; /* May be NULL. */
423 char *locale; /* May be NULL. */
425 /* Source information. */
426 char *dataset; /* May be NULL. */
427 char *datafile; /* May be NULL. */
428 time_t date; /* May be 0 if unknown. */
431 struct pivot_footnote **footnotes;
432 size_t n_footnotes, allocated_footnotes;
435 struct pivot_value *title;
436 struct pivot_value *subtype; /* Same as spv_item's subtype. */
437 struct pivot_value *corner_text;
438 struct pivot_value *caption;
442 struct area_style areas[PIVOT_N_AREAS];
443 struct table_border_style borders[PIVOT_N_BORDERS];
446 struct pivot_dimension **dimensions;
449 /* Allocation of dimensions to rows, columns, and layers. */
450 struct pivot_axis axes[PIVOT_N_AXES];
452 struct hmap cells; /* Contains "struct pivot_cell"s. */
455 /* Creating and destroy pivot tables. */
456 struct pivot_table *pivot_table_create (const char *title);
457 struct pivot_table *pivot_table_create__ (struct pivot_value *title,
458 const char *subtype);
459 struct pivot_table *pivot_table_create_for_text (struct pivot_value *title,
460 struct pivot_value *content);
462 struct pivot_table *pivot_table_ref (const struct pivot_table *);
463 void pivot_table_unref (struct pivot_table *);
464 bool pivot_table_is_shared (const struct pivot_table *);
466 /* Format of PIVOT_RC_COUNT cells. */
467 void pivot_table_set_weight_var (struct pivot_table *,
468 const struct variable *);
469 void pivot_table_set_weight_format (struct pivot_table *,
470 const struct fmt_spec *);
473 bool pivot_table_is_empty (const struct pivot_table *);
476 void pivot_table_submit (struct pivot_table *);
479 void pivot_table_put (struct pivot_table *, const size_t *dindexes, size_t n,
480 struct pivot_value *);
481 void pivot_table_put1 (struct pivot_table *, size_t idx1,
482 struct pivot_value *);
483 void pivot_table_put2 (struct pivot_table *, size_t idx1, size_t idx2,
484 struct pivot_value *);
485 void pivot_table_put3 (struct pivot_table *, size_t idx1, size_t idx2,
486 size_t idx3, struct pivot_value *);
487 void pivot_table_put4 (struct pivot_table *, size_t idx1, size_t idx2,
488 size_t idx3, size_t idx4, struct pivot_value *);
490 const struct pivot_value *pivot_table_get (const struct pivot_table *,
491 const size_t *dindexes);
493 struct pivot_value *pivot_table_get_rw (struct pivot_table *,
494 const size_t *dindexes);
498 Use pivot_table_create_footnote() to create a footnote.
499 Use pivot_value_add_footnote() to add a reference to a footnote. */
500 struct pivot_footnote
503 struct pivot_value *content;
504 struct pivot_value *marker;
508 struct pivot_footnote *pivot_table_create_footnote (
509 struct pivot_table *, struct pivot_value *content);
510 struct pivot_footnote *pivot_table_create_footnote__ (
511 struct pivot_table *, size_t idx,
512 struct pivot_value *marker, struct pivot_value *content);
514 void pivot_footnote_destroy (struct pivot_footnote *);
517 void pivot_table_convert_indexes_ptod (const struct pivot_table *,
518 const size_t *pindexes[PIVOT_N_AXES],
520 size_t *pivot_table_enumerate_axis (const struct pivot_table *,
521 enum pivot_axis_type,
522 const size_t *layer_indexes,
523 bool omit_empty, size_t *n);
524 #define PIVOT_ENUMERATION_FOR_EACH(INDEXES, ENUMERATION, AXIS) \
525 for ((INDEXES) = (ENUMERATION); *(INDEXES) != SIZE_MAX; \
526 (INDEXES) += MAX (1, (AXIS)->n_dimensions))
528 void pivot_table_assign_label_depth (struct pivot_table *);
530 void pivot_table_dump (const struct pivot_table *, int indentation);
534 enum pivot_value_type
536 PIVOT_VALUE_NUMERIC, /* A value of a numeric variable. */
537 PIVOT_VALUE_STRING, /* A value of a string variable. */
538 PIVOT_VALUE_VARIABLE, /* Name of a variable. */
539 PIVOT_VALUE_TEXT, /* Text. */
540 PIVOT_VALUE_TEMPLATE, /* Templated text. */
543 /* A pivot_value is the content of a single pivot table cell. A pivot_value is
544 also a pivot table's title, caption, footnote marker and contents, and so
547 A given pivot_value is one of:
549 1. A number resulting from a calculation (PIVOT_VALUE_NUMERIC). Use
550 pivot_value_new_number() to create such a pivot_value.
552 A numeric pivot_value has an associated display format (usually an F or
553 PCT format). This format can be set directly on the pivot_value, but
554 that is not usually the easiest way. Instead, it is usually true that
555 all of the values in a single category should have the same format
556 (e.g. all "Significance" values might use format F40.3), so PSPP makes
557 it easy to set the default format for a category while creating the
558 category. See pivot_dimension_create() for more details.
560 For numbers that should be displayed as integers,
561 pivot_value_new_integer() can occasionally be a useful special case.
563 2. A numeric or string value obtained from data (PIVOT_VALUE_NUMERIC or
564 PIVOT_VALUE_STRING). If such a value corresponds to a variable, then the
565 variable's name can be attached to the pivot_value. If the value has a
566 value label, then that can also be attached. When a label is present,
567 the user can control whether to show the value or the label or both.
569 Use pivot_value_new_var_value() to create pivot_values of these kinds.
571 3. A variable name (PIVOT_VALUE_VARIABLE). The variable label, if any, can
572 be attached too, and again the user can control whether to show the value
573 or the label or both.
575 4. A text string (PIVOT_VALUE_TEXT). The value stores the string in English
576 and translated into the output language (localized). Use
577 pivot_value_new_text() or pivot_value_new_text_format() for those cases.
578 In some cases, only an English or a localized version is available for
579 one reason or another, although this is regrettable; in those cases, use
580 pivot_value_new_user_text() or pivot_value_new_user_text_nocopy().
582 (There is also a PIVOT_VALUE_TEMPLATE but PSPP does not yet create these
589 A pivot_value may reference any number of footnotes. Use
590 pivot_value_add_footnote() to add a footnote reference. The footnotes being
591 referenced must first be created with pivot_table_create_footnote().
597 A pivot_value can have specific font and cell styles. Only the user should
602 struct font_style *font_style;
603 struct cell_style *cell_style;
610 const struct pivot_footnote **footnotes;
613 enum pivot_value_type type;
616 /* PIVOT_VALUE_NUMERIC. */
619 double x; /* The numeric value. */
620 struct fmt_spec format; /* Format to display 'x'. */
621 char *var_name; /* May be NULL. */
622 char *value_label; /* May be NULL. */
623 enum settings_value_show show; /* Show value or label or both? */
627 /* PIVOT_VALUE_STRING. */
630 char *s; /* The string value. */
631 bool hex; /* Display in hex? */
632 char *var_name; /* May be NULL. */
633 char *value_label; /* May be NULL. */
634 enum settings_value_show show; /* Show value or label or both? */
638 /* PIVOT_VALUE_VARIABLE. */
642 char *var_label; /* May be NULL. */
643 enum settings_value_show show; /* Show name or label or both? */
647 /* PIVOT_VALUE_TEXT. */
650 char *local; /* Localized. */
651 char *c; /* English. */
652 char *id; /* Identifier. */
657 /* PIVOT_VALUE_TEMPLATE. */
660 char *local; /* Localized. */
661 char *id; /* Identifier. */
662 struct pivot_argument *args;
669 /* Numbers resulting from calculations. */
670 struct pivot_value *pivot_value_new_number (double);
671 struct pivot_value *pivot_value_new_integer (double);
673 /* Values from data. */
674 struct pivot_value *pivot_value_new_var_value (
675 const struct variable *, const union value *);
676 struct pivot_value *pivot_value_new_value (const union value *, int width,
677 const struct fmt_spec *,
678 const char *encoding);
680 /* Values from variable names. */
681 struct pivot_value *pivot_value_new_variable (const struct variable *);
683 /* Values from text strings. */
684 struct pivot_value *pivot_value_new_text (const char *);
685 struct pivot_value *pivot_value_new_text_format (const char *, ...)
686 PRINTF_FORMAT (1, 2);
688 struct pivot_value *pivot_value_new_user_text (const char *, size_t length);
689 struct pivot_value *pivot_value_new_user_text_nocopy (char *);
692 void pivot_value_add_footnote (struct pivot_value *, const struct pivot_footnote *);
694 /* Numeric formats. */
695 void pivot_value_set_rc (const struct pivot_table *, struct pivot_value *,
698 /* Converting a pivot_value to a string for display. */
699 char *pivot_value_to_string (const struct pivot_value *,
700 enum settings_value_show show_values,
701 enum settings_value_show show_variables);
702 void pivot_value_format (const struct pivot_value *,
703 enum settings_value_show show_values,
704 enum settings_value_show show_variables,
706 bool pivot_value_format_body (const struct pivot_value *,
707 enum settings_value_show show_values,
708 enum settings_value_show show_variables,
711 void pivot_value_destroy (struct pivot_value *);
714 void pivot_value_get_style (struct pivot_value *,
715 const struct font_style *base_font_style,
716 const struct cell_style *base_cell_style,
717 struct area_style *);
718 void pivot_value_set_style (struct pivot_value *, const struct area_style *);
720 /* Template arguments. */
721 struct pivot_argument
724 struct pivot_value **values;
727 void pivot_argument_uninit (struct pivot_argument *);
729 /* One piece of data within a pivot table. */
732 struct hmap_node hmap_node; /* In struct pivot_table's 'cells' hmap. */
733 struct pivot_value *value;
734 unsigned int idx[]; /* One index per table dimension. */
737 #endif /* output/pivot-table.h */