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);
111 /* Table borders for styling purposes. */
117 PIVOT_BORDER_OUTER_LEFT,
118 PIVOT_BORDER_OUTER_TOP,
119 PIVOT_BORDER_OUTER_RIGHT,
120 PIVOT_BORDER_OUTER_BOTTOM,
123 PIVOT_BORDER_INNER_LEFT,
124 PIVOT_BORDER_INNER_TOP,
125 PIVOT_BORDER_INNER_RIGHT,
126 PIVOT_BORDER_INNER_BOTTOM,
129 PIVOT_BORDER_DATA_LEFT,
130 PIVOT_BORDER_DATA_TOP,
133 PIVOT_BORDER_DIM_ROW_HORZ,
134 PIVOT_BORDER_DIM_ROW_VERT,
135 PIVOT_BORDER_DIM_COL_HORZ,
136 PIVOT_BORDER_DIM_COL_VERT,
139 PIVOT_BORDER_CAT_ROW_HORZ,
140 PIVOT_BORDER_CAT_ROW_VERT,
141 PIVOT_BORDER_CAT_COL_HORZ,
142 PIVOT_BORDER_CAT_COL_VERT,
147 const char *pivot_border_to_string (enum pivot_border);
149 /* Sizing for rows or columns of a rendered table. The comments below talk
150 about columns and their widths but they apply equally to rows and their
152 struct pivot_table_sizing
154 /* Specific column widths, in 1/96" units. */
158 /* Specific page breaks: 0-based columns after which a page break must
159 occur, e.g. a value of 1 requests a break after the second column. */
163 /* Keeps: columns to keep together on a page if possible. */
164 struct pivot_keep *keeps;
168 void pivot_table_sizing_uninit (struct pivot_table_sizing *);
170 /* A set of columns to keep together on a page if possible, e.g. ofs=1, n=10
171 requests keeping together the 2nd through 11th columns. */
174 size_t ofs; /* 0-based first column. */
175 size_t n; /* Number of columns. */
189 const char *pivot_axis_type_to_string (enum pivot_axis_type);
191 /* An axis within a pivot table. */
194 /* dimensions[0] is the innermost dimension,
195 dimensions[1] is the next outer dimension,
197 dimensions[n_dimensions - 1] is the outermost dimension. */
198 struct pivot_dimension **dimensions;
201 /* The number of rows or columns along the axis,
202 that is, the product of dimension[*]->n_leaves.
203 It is 0 if any dimension has 0 leaves. */
206 /* Sum of dimensions[*]->label_depth. */
210 /* Successively assigns to INDEXES (which should be a "size_t *") each of the
211 combinations of the categories in AXIS's dimensions, in lexicographic order
212 with the innermost dimension iterating most quickly.
214 The value assigned to INDEXES is dynamically allocated. If the client
215 breaks out of the loop prematurely, it needs to free it with free(). */
216 #define PIVOT_AXIS_FOR_EACH(INDEXES, AXIS) \
217 for ((INDEXES) = NULL; \
218 ((INDEXES) = pivot_axis_iterator_next (INDEXES, AXIS)) != NULL;)
219 size_t *pivot_axis_iterator_next (size_t *indexes, const struct pivot_axis *);
223 A pivot_dimension identifies the categories associated with a single
224 dimension within a multidimensional pivot table.
226 A dimension contains a collection of categories, which are the leaves in a
229 (A dimension or a group can contain zero categories, but this is unusual.
230 If a dimension contains no categories, then its table cannot contain any
233 struct pivot_dimension
235 /* table->axes[axis_type]->dimensions[level] == dimension. */
236 struct pivot_table *table;
237 enum pivot_axis_type axis_type;
238 size_t level; /* 0 for innermost dimension within axis. */
240 /* table->dimensions[top_index] == dimension. */
243 /* Hierarchy of categories within the dimension. The groups and categories
244 are sorted in the order that should be used for display. This might be
245 different from the original order produced for output if the user
248 The root must always be a group, although it is allowed to have no
250 struct pivot_category *root;
252 /* All of the leaves reachable via the root.
254 The indexing for presentation_leaves is presentation order, thus
255 presentation_leaves[i]->presentation_index == i. This order is the same
256 as would be produced by an in-order traversal of the groups. It is the
257 order into which the user reordered or sorted the categories.
259 The indexing for data_leaves is that used for idx[] in struct
260 pivot_cell, thus data_leaves[i]->data_index == i. This might differ
261 from what an in-order traversal of 'root' would yield, if the user
262 reordered categories. */
263 struct pivot_category **data_leaves;
264 struct pivot_category **presentation_leaves;
265 size_t n_leaves, allocated_leaves;
268 bool hide_all_labels;
270 /* Number of rows or columns needed to express the labels. */
274 struct pivot_dimension *pivot_dimension_create (
275 struct pivot_table *, enum pivot_axis_type, const char *name, ...)
277 #define pivot_dimension_create(...) \
278 pivot_dimension_create(__VA_ARGS__, NULL_SENTINEL)
279 struct pivot_dimension *pivot_dimension_create__ (struct pivot_table *,
280 enum pivot_axis_type,
281 struct pivot_value *name);
283 void pivot_dimension_destroy (struct pivot_dimension *);
285 void pivot_dimension_dump (const struct pivot_dimension *, int indentation);
287 /* A pivot_category is a leaf (a category) or a group:
289 - For a leaf, neither index is SIZE_MAX.
291 - For a group, both indexes are SIZE_MAX.
293 Do not use 'subs' or 'n_subs' to determine whether a category is a group,
294 because a group may (pathologically) have no leaves. */
295 struct pivot_category
297 struct pivot_value *name;
298 struct pivot_category *parent;
299 struct pivot_dimension *dimension;
300 size_t label_depth, extra_depth;
304 If show_label is true, then the group itself has a row (or a column)
305 giving the group's name. Otherwise, the group's own name is not
307 struct pivot_category **subs; /* Child categories or groups. */
308 size_t n_subs, allocated_subs;
309 bool show_label; /* Display a label for the group itself? */
310 bool show_label_in_corner;
313 struct fmt_spec format;
314 size_t group_index; /* In ->parent->subs[]. */
315 size_t data_index; /* In ->dimension->data_leaves[]. */
316 size_t presentation_index; /* In ->dimension->presentation_leaves[]. */
320 pivot_category_is_group (const struct pivot_category *category)
322 return category->data_index == SIZE_MAX;
326 pivot_category_is_leaf (const struct pivot_category *category)
328 return !pivot_category_is_group (category);
331 /* Creating leaf categories. */
332 int pivot_category_create_leaves (struct pivot_category *parent, ...)
334 #define pivot_category_create_leaves(...) \
335 pivot_category_create_leaves(__VA_ARGS__, NULL_SENTINEL)
337 int pivot_category_create_leaf (
338 struct pivot_category *parent, struct pivot_value *name);
339 int pivot_category_create_leaf_rc (
340 struct pivot_category *parent, struct pivot_value *name, const char *rc);
342 /* Creating category groups. */
343 struct pivot_category *pivot_category_create_group (
344 struct pivot_category *parent, const char *name, ...) SENTINEL (0);
345 #define pivot_category_create_group(...) \
346 pivot_category_create_group(__VA_ARGS__, NULL_SENTINEL)
347 struct pivot_category *pivot_category_create_group__ (
348 struct pivot_category *parent, struct pivot_value *name);
350 void pivot_category_destroy (struct pivot_category *);
352 /* Pivot result classes.
354 These are used to mark leaf categories as having particular types of data,
355 to set their numeric formats. The formats that actually get used for these
356 classes are in the result_classes[] global array in pivot-table.c, except
357 that PIVOT_RC_OTHER comes from settings_get_format() and PIVOT_RC_COUNT
358 should come from the weight variable in the dataset's dictionary. */
359 #define PIVOT_RC_OTHER ("RC_OTHER")
360 #define PIVOT_RC_INTEGER ("RC_INTEGER")
361 #define PIVOT_RC_CORRELATION ("RC_CORRELATIONS")
362 #define PIVOT_RC_SIGNIFICANCE ("RC_SIGNIFICANCE")
363 #define PIVOT_RC_PERCENT ("RC_PERCENT")
364 #define PIVOT_RC_RESIDUAL ("RC_RESIDUAL")
365 #define PIVOT_RC_COUNT ("RC_COUNT")
367 bool pivot_result_class_change (const char *, const struct fmt_spec *);
369 /* Styling for a pivot table.
371 The division between this and the style information in struct pivot_table
372 seems fairly arbitrary. The ultimate reason for the division is simply
373 because that's how SPSS documentation and file formats do it. */
374 struct pivot_table_look
376 /* Reference count. A pivot_table_look may be shared between multiple
377 owners, indicated by a reference count greater than 1. When this is the
378 case, the pivot_table must not be modified. */
381 char *name; /* May be null. */
383 /* General properties. */
385 bool row_labels_in_corner;
386 int width_ranges[TABLE_N_AXES][2]; /* In 1/96" units. */
388 /* Footnote display settings. */
389 bool show_numeric_markers;
390 bool footnote_marker_superscripts;
393 struct table_area_style areas[PIVOT_N_AREAS];
394 struct table_border_style borders[PIVOT_N_BORDERS];
396 /* Print settings. */
397 bool print_all_layers;
398 bool paginate_layers;
399 bool shrink_to_fit[TABLE_N_AXES];
400 bool top_continuation, bottom_continuation;
402 size_t n_orphan_lines;
405 const struct pivot_table_look *pivot_table_look_builtin_default (void);
406 struct pivot_table_look *pivot_table_look_new_builtin_default (void);
407 struct pivot_table_look *pivot_table_look_ref (
408 const struct pivot_table_look *);
409 void pivot_table_look_unref (struct pivot_table_look *);
410 struct pivot_table_look *pivot_table_look_unshare (struct pivot_table_look *);
412 /* A pivot table. See the top of this file for more information. */
415 /* Reference count. A pivot_table may be shared between multiple owners,
416 indicated by a reference count greater than 1. When this is the case,
417 the pivot_table must not be modified. */
421 struct pivot_table_look *look;
423 /* Display settings. */
424 bool rotate_inner_column_labels;
425 bool rotate_outer_row_labels;
426 bool show_grid_lines;
428 size_t *current_layer; /* axis[PIVOT_AXIS_LAYER].n_dimensions elements. */
429 enum settings_value_show show_values;
430 enum settings_value_show show_variables;
431 struct fmt_spec weight_format;
433 /* Column and row sizing and page breaks.
434 sizing[TABLE_HORZ] is for columns, sizing[TABLE_VERT] is for rows. */
435 struct pivot_table_sizing sizing[TABLE_N_AXES];
437 /* Format settings. */
439 char decimal; /* Usually ',' or '.'. */
440 char grouping; /* Usually '.' or ','. */
441 char *ccs[5]; /* Custom currency. */
444 /* Command information. */
445 char *command_local; /* May be NULL. */
446 char *command_c; /* May be NULL. */
447 char *language; /* May be NULL. */
448 char *locale; /* May be NULL. */
450 /* Source information. */
451 char *dataset; /* May be NULL. */
452 char *datafile; /* May be NULL. */
453 time_t date; /* May be 0 if unknown. */
456 struct pivot_footnote **footnotes;
457 size_t n_footnotes, allocated_footnotes;
460 struct pivot_value *title;
461 struct pivot_value *subtype; /* Same as spv_item's subtype. */
462 struct pivot_value *corner_text;
463 struct pivot_value *caption;
467 struct pivot_dimension **dimensions;
470 /* Allocation of dimensions to rows, columns, and layers. */
471 struct pivot_axis axes[PIVOT_N_AXES];
473 struct hmap cells; /* Contains "struct pivot_cell"s. */
476 /* Creating and destroy pivot tables. */
477 struct pivot_table *pivot_table_create (const char *title);
478 struct pivot_table *pivot_table_create__ (struct pivot_value *title,
479 const char *subtype);
480 struct pivot_table *pivot_table_create_for_text (struct pivot_value *title,
481 struct pivot_value *content);
483 struct pivot_table *pivot_table_ref (const struct pivot_table *);
484 void pivot_table_unref (struct pivot_table *);
485 bool pivot_table_is_shared (const struct pivot_table *);
488 const struct pivot_table_look *pivot_table_get_look (
489 const struct pivot_table *);
490 void pivot_table_set_look (struct pivot_table *,
491 const struct pivot_table_look *);
493 /* Format of PIVOT_RC_COUNT cells. */
494 void pivot_table_set_weight_var (struct pivot_table *,
495 const struct variable *);
496 void pivot_table_set_weight_format (struct pivot_table *,
497 const struct fmt_spec *);
500 bool pivot_table_is_empty (const struct pivot_table *);
503 void pivot_table_submit (struct pivot_table *);
506 void pivot_table_put (struct pivot_table *, const size_t *dindexes, size_t n,
507 struct pivot_value *);
508 void pivot_table_put1 (struct pivot_table *, size_t idx1,
509 struct pivot_value *);
510 void pivot_table_put2 (struct pivot_table *, size_t idx1, size_t idx2,
511 struct pivot_value *);
512 void pivot_table_put3 (struct pivot_table *, size_t idx1, size_t idx2,
513 size_t idx3, struct pivot_value *);
514 void pivot_table_put4 (struct pivot_table *, size_t idx1, size_t idx2,
515 size_t idx3, size_t idx4, struct pivot_value *);
517 const struct pivot_value *pivot_table_get (const struct pivot_table *,
518 const size_t *dindexes);
520 struct pivot_value *pivot_table_get_rw (struct pivot_table *,
521 const size_t *dindexes);
525 Use pivot_table_create_footnote() to create a footnote.
526 Use pivot_value_add_footnote() to add a reference to a footnote. */
527 struct pivot_footnote
530 struct pivot_value *content;
531 struct pivot_value *marker;
535 struct pivot_footnote *pivot_table_create_footnote (
536 struct pivot_table *, struct pivot_value *content);
537 struct pivot_footnote *pivot_table_create_footnote__ (
538 struct pivot_table *, size_t idx,
539 struct pivot_value *marker, struct pivot_value *content);
541 void pivot_footnote_destroy (struct pivot_footnote *);
544 void pivot_table_convert_indexes_ptod (const struct pivot_table *,
545 const size_t *pindexes[PIVOT_N_AXES],
547 size_t *pivot_table_enumerate_axis (const struct pivot_table *,
548 enum pivot_axis_type,
549 const size_t *layer_indexes,
550 bool omit_empty, size_t *n);
551 #define PIVOT_ENUMERATION_FOR_EACH(INDEXES, ENUMERATION, AXIS) \
552 for ((INDEXES) = (ENUMERATION); *(INDEXES) != SIZE_MAX; \
553 (INDEXES) += MAX (1, (AXIS)->n_dimensions))
555 void pivot_table_assign_label_depth (struct pivot_table *);
557 void pivot_table_dump (const struct pivot_table *, int indentation);
561 enum pivot_value_type
563 PIVOT_VALUE_NUMERIC, /* A value of a numeric variable. */
564 PIVOT_VALUE_STRING, /* A value of a string variable. */
565 PIVOT_VALUE_VARIABLE, /* Name of a variable. */
566 PIVOT_VALUE_TEXT, /* Text. */
567 PIVOT_VALUE_TEMPLATE, /* Templated text. */
570 /* A pivot_value is the content of a single pivot table cell. A pivot_value is
571 also a pivot table's title, caption, footnote marker and contents, and so
574 A given pivot_value is one of:
576 1. A number resulting from a calculation (PIVOT_VALUE_NUMERIC). Use
577 pivot_value_new_number() to create such a pivot_value.
579 A numeric pivot_value has an associated display format (usually an F or
580 PCT format). This format can be set directly on the pivot_value, but
581 that is not usually the easiest way. Instead, it is usually true that
582 all of the values in a single category should have the same format
583 (e.g. all "Significance" values might use format F40.3), so PSPP makes
584 it easy to set the default format for a category while creating the
585 category. See pivot_dimension_create() for more details.
587 For numbers that should be displayed as integers,
588 pivot_value_new_integer() can occasionally be a useful special case.
590 2. A numeric or string value obtained from data (PIVOT_VALUE_NUMERIC or
591 PIVOT_VALUE_STRING). If such a value corresponds to a variable, then the
592 variable's name can be attached to the pivot_value. If the value has a
593 value label, then that can also be attached. When a label is present,
594 the user can control whether to show the value or the label or both.
596 Use pivot_value_new_var_value() to create pivot_values of these kinds.
598 3. A variable name (PIVOT_VALUE_VARIABLE). The variable label, if any, can
599 be attached too, and again the user can control whether to show the value
600 or the label or both.
602 4. A text string (PIVOT_VALUE_TEXT). The value stores the string in English
603 and translated into the output language (localized). Use
604 pivot_value_new_text() or pivot_value_new_text_format() for those cases.
605 In some cases, only an English or a localized version is available for
606 one reason or another, although this is regrettable; in those cases, use
607 pivot_value_new_user_text() or pivot_value_new_user_text_nocopy().
609 (There is also a PIVOT_VALUE_TEMPLATE but PSPP does not yet create these
616 A pivot_value may reference any number of footnotes. Use
617 pivot_value_add_footnote() to add a footnote reference. The footnotes being
618 referenced must first be created with pivot_table_create_footnote().
624 A pivot_value can have specific font and cell styles. Only the user should
629 struct font_style *font_style;
630 struct cell_style *cell_style;
637 const struct pivot_footnote **footnotes;
640 enum pivot_value_type type;
643 /* PIVOT_VALUE_NUMERIC. */
646 double x; /* The numeric value. */
647 struct fmt_spec format; /* Format to display 'x'. */
648 char *var_name; /* May be NULL. */
649 char *value_label; /* May be NULL. */
650 enum settings_value_show show; /* Show value or label or both? */
654 /* PIVOT_VALUE_STRING. */
657 char *s; /* The string value. */
658 bool hex; /* Display in hex? */
659 char *var_name; /* May be NULL. */
660 char *value_label; /* May be NULL. */
661 enum settings_value_show show; /* Show value or label or both? */
665 /* PIVOT_VALUE_VARIABLE. */
669 char *var_label; /* May be NULL. */
670 enum settings_value_show show; /* Show name or label or both? */
674 /* PIVOT_VALUE_TEXT. */
677 char *local; /* Localized. */
678 char *c; /* English. */
679 char *id; /* Identifier. */
684 /* PIVOT_VALUE_TEMPLATE. */
687 char *local; /* Localized. */
688 char *id; /* Identifier. */
689 struct pivot_argument *args;
696 /* Numbers resulting from calculations. */
697 struct pivot_value *pivot_value_new_number (double);
698 struct pivot_value *pivot_value_new_integer (double);
700 /* Values from data. */
701 struct pivot_value *pivot_value_new_var_value (
702 const struct variable *, const union value *);
703 struct pivot_value *pivot_value_new_value (const union value *, int width,
704 const struct fmt_spec *,
705 const char *encoding);
707 /* Values from variable names. */
708 struct pivot_value *pivot_value_new_variable (const struct variable *);
710 /* Values from text strings. */
711 struct pivot_value *pivot_value_new_text (const char *);
712 struct pivot_value *pivot_value_new_text_format (const char *, ...)
713 #if defined(__GNUC__) && ((__GNUC__ == 4 && __GNUC_MINOR__>= 4) || __GNUC__ > 4)
714 __attribute__((format(gnu_printf, 1, 2)));
716 __attribute__((format(__printf__, 1, 2)));
719 struct pivot_value *pivot_value_new_user_text (const char *, size_t length);
720 struct pivot_value *pivot_value_new_user_text_nocopy (char *);
723 void pivot_value_add_footnote (struct pivot_value *, const struct pivot_footnote *);
725 /* Numeric formats. */
726 void pivot_value_set_rc (const struct pivot_table *, struct pivot_value *,
729 /* Converting a pivot_value to a string for display. */
730 char *pivot_value_to_string (const struct pivot_value *,
731 enum settings_value_show show_values,
732 enum settings_value_show show_variables);
733 void pivot_value_format (const struct pivot_value *,
734 enum settings_value_show show_values,
735 enum settings_value_show show_variables,
737 bool pivot_value_format_body (const struct pivot_value *,
738 enum settings_value_show show_values,
739 enum settings_value_show show_variables,
742 void pivot_value_destroy (struct pivot_value *);
745 void pivot_value_get_style (struct pivot_value *,
746 const struct font_style *base_font_style,
747 const struct cell_style *base_cell_style,
748 struct table_area_style *);
749 void pivot_value_set_style (struct pivot_value *,
750 const struct table_area_style *);
752 /* Template arguments. */
753 struct pivot_argument
756 struct pivot_value **values;
759 void pivot_argument_uninit (struct pivot_argument *);
761 /* One piece of data within a pivot table. */
764 struct hmap_node hmap_node; /* In struct pivot_table's 'cells' hmap. */
765 struct pivot_value *value;
766 unsigned int idx[]; /* One index per table dimension. */
769 #endif /* output/pivot-table.h */