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.
64 2. Optionally, set the format to use for "count" values with
65 pivot_table_set_weight_var() or pivot_table_set_weight_format().
67 3. Create each dimension with pivot_dimension_create() and populate it with
68 categories and, possibly, with groups that contain the categories. This
69 call also assigns the dimension to an axis.
71 In simple cases, only a call to pivot_dimension_create() is needed.
72 Other functions such as pivot_category_create_group() can be used for
73 hierarchies of categories.
75 Sometimes it's easier to create categories in tandem with inserting data,
76 for example by adding a category for a variable just before inserting the
77 first cell for that variable. In that case, creating categories and
78 inserting data can be interleaved.
80 4. Insert data. For each cell, supply the category indexes, which are
81 assigned starting from 0 in the order in which the categories were
82 created in step 2, and the value to go in the cell. If the table has a
83 small, fixed number of dimensions, functions like, e.g.
84 pivot_table_put3() for 3 dimensions, can be used. The general function
85 pivot_table_put() works for other cases.
87 5. Output the table for user consumption. Use pivot_table_submit(). */
89 /* Pivot table display styling. */
91 /* Areas of a pivot table for styling purposes. */
96 PIVOT_AREA_FOOTER, /* Footnotes. */
97 PIVOT_AREA_CORNER, /* Top-left corner. */
98 PIVOT_AREA_COLUMN_LABELS,
99 PIVOT_AREA_ROW_LABELS,
101 PIVOT_AREA_LAYERS, /* Layer indication. */
105 const char *pivot_area_to_string (enum pivot_area);
107 /* Table borders for styling purposes. */
113 PIVOT_BORDER_OUTER_LEFT,
114 PIVOT_BORDER_OUTER_TOP,
115 PIVOT_BORDER_OUTER_RIGHT,
116 PIVOT_BORDER_OUTER_BOTTOM,
119 PIVOT_BORDER_INNER_LEFT,
120 PIVOT_BORDER_INNER_TOP,
121 PIVOT_BORDER_INNER_RIGHT,
122 PIVOT_BORDER_INNER_BOTTOM,
125 PIVOT_BORDER_DATA_LEFT,
126 PIVOT_BORDER_DATA_TOP,
129 PIVOT_BORDER_DIM_ROW_HORZ,
130 PIVOT_BORDER_DIM_ROW_VERT,
131 PIVOT_BORDER_DIM_COL_HORZ,
132 PIVOT_BORDER_DIM_COL_VERT,
135 PIVOT_BORDER_CAT_ROW_HORZ,
136 PIVOT_BORDER_CAT_ROW_VERT,
137 PIVOT_BORDER_CAT_COL_HORZ,
138 PIVOT_BORDER_CAT_COL_VERT,
143 const char *pivot_border_to_string (enum pivot_border);
145 /* Sizing for rows or columns of a rendered table. The comments below talk
146 about columns and their widths but they apply equally to rows and their
148 struct pivot_table_sizing
150 /* Specific column widths, in 1/96" units. */
154 /* Specific page breaks: 0-based columns after which a page break must
155 occur, e.g. a value of 1 requests a break after the second column. */
159 /* Keeps: columns to keep together on a page if possible. */
160 struct pivot_keep *keeps;
164 void pivot_table_sizing_uninit (struct pivot_table_sizing *);
166 /* A set of columns to keep together on a page if possible, e.g. ofs=1, n=10
167 requests keeping together the 2nd through 11th columns. */
170 size_t ofs; /* 0-based first column. */
171 size_t n; /* Number of columns. */
185 const char *pivot_axis_type_to_string (enum pivot_axis_type);
187 /* An axis within a pivot table. */
190 /* dimensions[0] is the innermost dimension,
191 dimensions[1] is the next outer dimension,
193 dimensions[n_dimensions - 1] is the outermost dimension. */
194 struct pivot_dimension **dimensions;
197 /* The number of rows or columns along the axis,
198 that is, the product of dimension[*]->n_leaves.
199 It is 0 if any dimension has 0 leaves. */
202 /* Sum of dimensions[*]->label_depth. */
206 /* Successively assigns to INDEXES (which should be a "size_t *") each of the
207 combinations of the categories in AXIS's dimensions, in lexicographic order
208 with the innermost dimension iterating most quickly.
210 The value assigned to INDEXES is dynamically allocated. If the client
211 breaks out of the loop prematurely, it needs to free it with free(). */
212 #define PIVOT_AXIS_FOR_EACH(INDEXES, AXIS) \
213 for ((INDEXES) = NULL; \
214 ((INDEXES) = pivot_axis_iterator_next (INDEXES, AXIS)) != NULL;)
215 size_t *pivot_axis_iterator_next (size_t *indexes, const struct pivot_axis *);
219 A pivot_dimension identifies the categories associated with a single
220 dimension within a multidimensional pivot table.
222 A dimension contains a collection of categories, which are the leaves in a
225 (A dimension or a group can contain zero categories, but this is unusual.
226 If a dimension contains no categories, then its table cannot contain any
229 struct pivot_dimension
231 /* table->axes[axis_type]->dimensions[level] == dimension. */
232 struct pivot_table *table;
233 enum pivot_axis_type axis_type;
234 size_t level; /* 0 for innermost dimension within axis. */
236 /* table->dimensions[top_index] == dimension. */
239 /* Hierarchy of categories within the dimension. The groups and categories
240 are sorted in the order that should be used for display. This might be
241 different from the original order produced for output if the user
244 The root must always be a group, although it is allowed to have no
246 struct pivot_category *root;
248 /* All of the leaves reachable via the root.
250 The indexing for presentation_leaves is presentation order, thus
251 presentation_leaves[i]->presentation_index == i. This order is the same
252 as would be produced by an in-order traversal of the groups. It is the
253 order into which the user reordered or sorted the categories.
255 The indexing for data_leaves is that used for idx[] in struct
256 pivot_cell, thus data_leaves[i]->data_index == i. This might differ
257 from what an in-order traversal of 'root' would yield, if the user
258 reordered categories. */
259 struct pivot_category **data_leaves;
260 struct pivot_category **presentation_leaves;
261 size_t n_leaves, allocated_leaves;
264 bool hide_all_labels;
266 /* Number of rows or columns needed to express the labels. */
270 struct pivot_dimension *pivot_dimension_create (
271 struct pivot_table *, enum pivot_axis_type, const char *name, ...)
273 #define pivot_dimension_create(...) \
274 pivot_dimension_create(__VA_ARGS__, NULL_SENTINEL)
275 struct pivot_dimension *pivot_dimension_create__ (struct pivot_table *,
276 enum pivot_axis_type,
277 struct pivot_value *name);
279 void pivot_dimension_destroy (struct pivot_dimension *);
281 void pivot_dimension_dump (const struct pivot_dimension *, int indentation);
283 /* A pivot_category is a leaf (a category) or a group:
285 - For a leaf, neither index is SIZE_MAX.
287 - For a group, both indexes are SIZE_MAX.
289 Do not use 'subs' or 'n_subs' to determine whether a category is a group,
290 because a group may (pathologically) have no leaves. */
291 struct pivot_category
293 struct pivot_value *name;
294 struct pivot_category *parent;
295 struct pivot_dimension *dimension;
296 size_t label_depth, extra_depth;
300 If show_label is true, then the group itself has a row (or a column)
301 giving the group's name. Otherwise, the group's own name is not
303 struct pivot_category **subs; /* Child categories or groups. */
304 size_t n_subs, allocated_subs;
305 bool show_label; /* Display a label for the group itself? */
306 bool show_label_in_corner;
309 struct fmt_spec format;
310 size_t group_index; /* In ->parent->subs[]. */
311 size_t data_index; /* In ->dimension->data_leaves[]. */
312 size_t presentation_index; /* In ->dimension->presentation_leaves[]. */
316 pivot_category_is_group (const struct pivot_category *category)
318 return category->data_index == SIZE_MAX;
322 pivot_category_is_leaf (const struct pivot_category *category)
324 return !pivot_category_is_group (category);
327 /* Creating leaf categories. */
328 int pivot_category_create_leaves (struct pivot_category *parent, ...)
330 #define pivot_category_create_leaves(...) \
331 pivot_category_create_leaves(__VA_ARGS__, NULL_SENTINEL)
333 int pivot_category_create_leaf (
334 struct pivot_category *parent, struct pivot_value *name);
335 int pivot_category_create_leaf_rc (
336 struct pivot_category *parent, struct pivot_value *name, const char *rc);
338 /* Creating category groups. */
339 struct pivot_category *pivot_category_create_group (
340 struct pivot_category *parent, const char *name, ...) SENTINEL (0);
341 #define pivot_category_create_group(...) \
342 pivot_category_create_group(__VA_ARGS__, NULL_SENTINEL)
343 struct pivot_category *pivot_category_create_group__ (
344 struct pivot_category *parent, struct pivot_value *name);
346 void pivot_category_destroy (struct pivot_category *);
348 /* Pivot result classes.
350 These are used to mark leaf categories as having particular types of data,
351 to set their numeric formats. The formats that actually get used for these
352 classes are in the result_classes[] global array in pivot-table.c, except
353 that PIVOT_RC_OTHER comes from settings_get_format() and PIVOT_RC_COUNT
354 should come from the weight variable in the dataset's dictionary. */
355 #define PIVOT_RC_OTHER ("RC_OTHER")
356 #define PIVOT_RC_INTEGER ("RC_INTEGER")
357 #define PIVOT_RC_CORRELATION ("RC_CORRELATIONS")
358 #define PIVOT_RC_SIGNIFICANCE ("RC_SIGNIFICANCE")
359 #define PIVOT_RC_PERCENT ("RC_PERCENT")
360 #define PIVOT_RC_RESIDUAL ("RC_RESIDUAL")
361 #define PIVOT_RC_COUNT ("RC_COUNT")
363 bool pivot_result_class_change (const char *, const struct fmt_spec *);
365 /* Styling for a pivot table.
367 The division between this and the style information in struct pivot_table
368 seems fairly arbitrary. The ultimate reason for the division is simply
369 because that's how SPSS documentation and file formats do it. */
370 struct pivot_table_look
372 /* Reference count. A pivot_table_look may be shared between multiple
373 owners, indicated by a reference count greater than 1. When this is the
374 case, the pivot_table must not be modified. */
377 char *name; /* May be null. */
379 /* General properties. */
381 bool row_labels_in_corner;
382 int width_ranges[TABLE_N_AXES][2]; /* In 1/96" units. */
384 /* Footnote display settings. */
385 bool show_numeric_markers;
386 bool footnote_marker_superscripts;
389 struct table_area_style areas[PIVOT_N_AREAS];
390 struct table_border_style borders[PIVOT_N_BORDERS];
392 /* Print settings. */
393 bool print_all_layers;
394 bool paginate_layers;
395 bool shrink_to_fit[TABLE_N_AXES];
396 bool top_continuation, bottom_continuation;
398 size_t n_orphan_lines;
401 const struct pivot_table_look *pivot_table_look_get_default (void);
402 void pivot_table_look_set_default (const struct pivot_table_look *);
404 char *pivot_table_look_read (const char *, struct pivot_table_look **)
407 const struct pivot_table_look *pivot_table_look_builtin_default (void);
408 struct pivot_table_look *pivot_table_look_new_builtin_default (void);
409 struct pivot_table_look *pivot_table_look_ref (
410 const struct pivot_table_look *);
411 void pivot_table_look_unref (struct pivot_table_look *);
412 struct pivot_table_look *pivot_table_look_unshare (struct pivot_table_look *);
414 /* A pivot table. See the top of this file for more information. */
417 /* Reference count. A pivot_table may be shared between multiple owners,
418 indicated by a reference count greater than 1. When this is the case,
419 the pivot_table must not be modified. */
423 struct pivot_table_look *look;
425 /* Display settings. */
426 bool rotate_inner_column_labels;
427 bool rotate_outer_row_labels;
428 bool show_grid_lines;
431 size_t *current_layer; /* axes[PIVOT_AXIS_LAYER].n_dimensions elements. */
432 enum settings_value_show show_values;
433 enum settings_value_show show_variables;
434 struct fmt_spec weight_format;
436 /* Column and row sizing and page breaks.
437 sizing[TABLE_HORZ] is for columns, sizing[TABLE_VERT] is for rows. */
438 struct pivot_table_sizing sizing[TABLE_N_AXES];
440 /* Format settings. */
442 char decimal; /* Usually ',' or '.'. */
443 char grouping; /* Usually '.' or ','. */
444 char *ccs[5]; /* Custom currency. */
447 /* Command information. */
448 char *command_local; /* May be NULL. */
449 char *command_c; /* May be NULL. */
450 char *language; /* May be NULL. */
451 char *locale; /* May be NULL. */
453 /* Source information. */
454 char *dataset; /* May be NULL. */
455 char *datafile; /* May be NULL. */
456 time_t date; /* May be 0 if unknown. */
459 struct pivot_footnote **footnotes;
460 size_t n_footnotes, allocated_footnotes;
463 struct pivot_value *title;
464 struct pivot_value *subtype; /* Same as spv_item's subtype. */
465 struct pivot_value *corner_text;
466 struct pivot_value *caption;
467 char *notes; /* Shown as tooltip. */
470 struct pivot_dimension **dimensions;
473 /* Allocation of dimensions to rows, columns, and layers. */
474 struct pivot_axis axes[PIVOT_N_AXES];
476 struct hmap cells; /* Contains "struct pivot_cell"s. */
479 /* Creating and destroy pivot tables. */
480 struct pivot_table *pivot_table_create (const char *title);
481 struct pivot_table *pivot_table_create__ (struct pivot_value *title,
482 const char *subtype);
483 struct pivot_table *pivot_table_create_for_text (struct pivot_value *title,
484 struct pivot_value *content);
486 struct pivot_table *pivot_table_ref (const struct pivot_table *);
487 void pivot_table_unref (struct pivot_table *);
488 bool pivot_table_is_shared (const struct pivot_table *);
491 const struct pivot_table_look *pivot_table_get_look (
492 const struct pivot_table *);
493 void pivot_table_set_look (struct pivot_table *,
494 const struct pivot_table_look *);
496 /* Format of PIVOT_RC_COUNT cells. */
497 void pivot_table_set_weight_var (struct pivot_table *,
498 const struct variable *);
499 void pivot_table_set_weight_format (struct pivot_table *,
500 const struct fmt_spec *);
503 bool pivot_table_is_empty (const struct pivot_table *);
506 void pivot_table_submit (struct pivot_table *);
509 void pivot_table_put (struct pivot_table *, const size_t *dindexes, size_t n,
510 struct pivot_value *);
511 void pivot_table_put1 (struct pivot_table *, size_t idx1,
512 struct pivot_value *);
513 void pivot_table_put2 (struct pivot_table *, size_t idx1, size_t idx2,
514 struct pivot_value *);
515 void pivot_table_put3 (struct pivot_table *, size_t idx1, size_t idx2,
516 size_t idx3, struct pivot_value *);
517 void pivot_table_put4 (struct pivot_table *, size_t idx1, size_t idx2,
518 size_t idx3, size_t idx4, struct pivot_value *);
520 const struct pivot_value *pivot_table_get (const struct pivot_table *,
521 const size_t *dindexes);
523 struct pivot_value *pivot_table_get_rw (struct pivot_table *,
524 const size_t *dindexes);
528 Use pivot_table_create_footnote() to create a footnote.
529 Use pivot_value_add_footnote() to add a reference to a footnote. */
530 struct pivot_footnote
533 struct pivot_value *content;
534 struct pivot_value *marker;
538 struct pivot_footnote *pivot_table_create_footnote (
539 struct pivot_table *, struct pivot_value *content);
540 struct pivot_footnote *pivot_table_create_footnote__ (
541 struct pivot_table *, size_t idx,
542 struct pivot_value *marker, struct pivot_value *content);
544 void pivot_footnote_destroy (struct pivot_footnote *);
547 void pivot_table_convert_indexes_ptod (const struct pivot_table *,
548 const size_t *pindexes[PIVOT_N_AXES],
550 size_t *pivot_table_enumerate_axis (const struct pivot_table *,
551 enum pivot_axis_type,
552 const size_t *layer_indexes,
553 bool omit_empty, size_t *n);
554 #define PIVOT_ENUMERATION_FOR_EACH(INDEXES, ENUMERATION, AXIS) \
555 for ((INDEXES) = (ENUMERATION); *(INDEXES) != SIZE_MAX; \
556 (INDEXES) += MAX (1, (AXIS)->n_dimensions))
558 void pivot_table_assign_label_depth (struct pivot_table *);
560 void pivot_table_dump (const struct pivot_table *, int indentation);
564 enum pivot_value_type
566 PIVOT_VALUE_NUMERIC, /* A value of a numeric variable. */
567 PIVOT_VALUE_STRING, /* A value of a string variable. */
568 PIVOT_VALUE_VARIABLE, /* Name of a variable. */
569 PIVOT_VALUE_TEXT, /* Text. */
570 PIVOT_VALUE_TEMPLATE, /* Templated text. */
573 /* A pivot_value is the content of a single pivot table cell. A pivot_value is
574 also a pivot table's title, caption, footnote marker and contents, and so
577 A given pivot_value is one of:
579 1. A number resulting from a calculation (PIVOT_VALUE_NUMERIC). Use
580 pivot_value_new_number() to create such a pivot_value.
582 A numeric pivot_value has an associated display format (usually an F or
583 PCT format). This format can be set directly on the pivot_value, but
584 that is not usually the easiest way. Instead, it is usually true that
585 all of the values in a single category should have the same format
586 (e.g. all "Significance" values might use format F40.3), so PSPP makes
587 it easy to set the default format for a category while creating the
588 category. See pivot_dimension_create() for more details.
590 For numbers that should be displayed as integers,
591 pivot_value_new_integer() can occasionally be a useful special case.
593 2. A numeric or string value obtained from data (PIVOT_VALUE_NUMERIC or
594 PIVOT_VALUE_STRING). If such a value corresponds to a variable, then the
595 variable's name can be attached to the pivot_value. If the value has a
596 value label, then that can also be attached. When a label is present,
597 the user can control whether to show the value or the label or both.
599 Use pivot_value_new_var_value() to create pivot_values of these kinds.
601 3. A variable name (PIVOT_VALUE_VARIABLE). The variable label, if any, can
602 be attached too, and again the user can control whether to show the value
603 or the label or both.
605 4. A text string (PIVOT_VALUE_TEXT). The value stores the string in English
606 and translated into the output language (localized). Use
607 pivot_value_new_text() or pivot_value_new_text_format() for those cases.
608 In some cases, only an English or a localized version is available for
609 one reason or another, although this is regrettable; in those cases, use
610 pivot_value_new_user_text() or pivot_value_new_user_text_nocopy().
612 (There is also a PIVOT_VALUE_TEMPLATE but PSPP does not yet create these
619 A pivot_value may reference any number of footnotes. Use
620 pivot_value_add_footnote() to add a footnote reference. The footnotes being
621 referenced must first be created with pivot_table_create_footnote().
627 A pivot_value can have specific font and cell styles. Only the user should
632 struct font_style *font_style;
633 struct cell_style *cell_style;
638 const struct pivot_footnote **footnotes;
641 enum pivot_value_type type;
644 /* PIVOT_VALUE_NUMERIC. */
647 double x; /* The numeric value. */
648 struct fmt_spec format; /* Format to display 'x'. */
649 char *var_name; /* May be NULL. */
650 char *value_label; /* May be NULL. */
651 enum settings_value_show show; /* Show value or label or both? */
655 /* PIVOT_VALUE_STRING. */
658 char *s; /* The string value. */
659 bool hex; /* Display in hex? */
660 char *var_name; /* May be NULL. */
661 char *value_label; /* May be NULL. */
662 enum settings_value_show show; /* Show value or label or both? */
666 /* PIVOT_VALUE_VARIABLE. */
670 char *var_label; /* May be NULL. */
671 enum settings_value_show show; /* Show name or label or both? */
675 /* PIVOT_VALUE_TEXT. */
678 char *local; /* Localized. */
679 char *c; /* English. */
680 char *id; /* Identifier. */
685 /* PIVOT_VALUE_TEMPLATE. */
688 char *local; /* Localized. */
689 char *id; /* Identifier. */
690 struct pivot_argument *args;
697 /* Numbers resulting from calculations. */
698 struct pivot_value *pivot_value_new_number (double);
699 struct pivot_value *pivot_value_new_integer (double);
701 /* Values from data. */
702 struct pivot_value *pivot_value_new_var_value (
703 const struct variable *, const union value *);
704 struct pivot_value *pivot_value_new_value (const union value *, int width,
705 const struct fmt_spec *,
706 const char *encoding);
708 /* Values from variable names. */
709 struct pivot_value *pivot_value_new_variable (const struct variable *);
711 /* Values from text strings. */
712 struct pivot_value *pivot_value_new_text (const char *);
713 struct pivot_value *pivot_value_new_text_format (const char *, ...)
714 #if defined(__GNUC__) && ((__GNUC__ == 4 && __GNUC_MINOR__>= 4) || __GNUC__ > 4)
715 __attribute__((format(gnu_printf, 1, 2)));
717 __attribute__((format(__printf__, 1, 2)));
720 struct pivot_value *pivot_value_new_user_text (const char *, size_t length);
721 struct pivot_value *pivot_value_new_user_text_nocopy (char *);
724 void pivot_value_add_footnote (struct pivot_value *, const struct pivot_footnote *);
726 /* Numeric formats. */
727 void pivot_value_set_rc (const struct pivot_table *, struct pivot_value *,
730 /* Converting a pivot_value to a string for display. */
731 char *pivot_value_to_string (const struct pivot_value *,
732 enum settings_value_show show_values,
733 enum settings_value_show show_variables);
734 void pivot_value_format (const struct pivot_value *,
735 enum settings_value_show show_values,
736 enum settings_value_show show_variables,
738 bool pivot_value_format_body (const struct pivot_value *,
739 enum settings_value_show show_values,
740 enum settings_value_show show_variables,
743 void pivot_value_destroy (struct pivot_value *);
746 void pivot_value_get_style (struct pivot_value *,
747 const struct font_style *base_font_style,
748 const struct cell_style *base_cell_style,
749 struct table_area_style *);
750 void pivot_value_set_style (struct pivot_value *,
751 const struct table_area_style *);
753 /* Template arguments. */
754 struct pivot_argument
757 struct pivot_value **values;
760 void pivot_argument_uninit (struct pivot_argument *);
762 /* One piece of data within a pivot table. */
765 struct hmap_node hmap_node; /* In struct pivot_table's 'cells' hmap. */
766 struct pivot_value *value;
767 unsigned int idx[]; /* One index per table dimension. */
770 #endif /* output/pivot-table.h */