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. Create each dimension with pivot_dimension_create() and populate it with
65 categories and, possibly, with groups that contain the categories. This
66 call also assigns the dimension to an axis.
68 In simple cases, only a call to pivot_dimension_create() is needed.
69 Other functions such as pivot__create_group() can be used for
70 hierarchies of categories.
72 Sometimes it's easier to create categories in tandem with inserting data,
73 for example by adding a category for a variable just before inserting the
74 first cell for that variable. In that case, creating categories and
75 inserting data can be interleaved.
77 3. Insert data. For each cell, supply the category indexes, which are
78 assigned starting from 0 in the order in which the categories were
79 created in step 2, and the value to go in the cell. If the table has a
80 small, fixed number of dimensions, functions like, e.g.
81 pivot_table_put3() for 3 dimensions, can be used. The general function
82 pivot_table_put() works for other cases.
84 4. Output the table for user consumption. Use pivot_table_submit(). */
86 /* Pivot table display styling. */
88 /* Areas of a pivot table for styling purposes. */
93 PIVOT_AREA_FOOTER, /* Footnotes. */
94 PIVOT_AREA_CORNER, /* Top-left corner. */
95 PIVOT_AREA_COLUMN_LABELS,
96 PIVOT_AREA_ROW_LABELS,
98 PIVOT_AREA_LAYERS, /* Layer indication. */
102 const char *pivot_area_to_string (enum pivot_area);
104 /* Table borders for styling purposes. */
110 PIVOT_BORDER_OUTER_LEFT,
111 PIVOT_BORDER_OUTER_TOP,
112 PIVOT_BORDER_OUTER_RIGHT,
113 PIVOT_BORDER_OUTER_BOTTOM,
116 PIVOT_BORDER_INNER_LEFT,
117 PIVOT_BORDER_INNER_TOP,
118 PIVOT_BORDER_INNER_RIGHT,
119 PIVOT_BORDER_INNER_BOTTOM,
122 PIVOT_BORDER_DATA_LEFT,
123 PIVOT_BORDER_DATA_TOP,
126 PIVOT_BORDER_DIM_ROW_HORZ,
127 PIVOT_BORDER_DIM_ROW_VERT,
128 PIVOT_BORDER_DIM_COL_HORZ,
129 PIVOT_BORDER_DIM_COL_VERT,
132 PIVOT_BORDER_CAT_ROW_HORZ,
133 PIVOT_BORDER_CAT_ROW_VERT,
134 PIVOT_BORDER_CAT_COL_HORZ,
135 PIVOT_BORDER_CAT_COL_VERT,
140 const char *pivot_border_to_string (enum pivot_border);
142 /* Sizing for rows or columns of a rendered table. The comments below talk
143 about columns and their widths but they apply equally to rows and their
145 struct pivot_table_sizing
147 /* Minimum and maximum column width, in 1/96" units. */
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 /* A pivot table. See the top of this file for more information. */
368 /* Display settings. */
369 bool rotate_inner_column_labels;
370 bool rotate_outer_row_labels;
371 bool row_labels_in_corner;
372 bool show_grid_lines;
374 size_t *current_layer; /* axis[PIVOT_AXIS_LAYER]->n_dimensions elements. */
376 enum settings_value_show show_values;
377 enum settings_value_show show_variables;
378 struct fmt_spec weight_format;
380 /* Footnote display settings. */
381 bool show_numeric_markers;
382 bool footnote_marker_superscripts;
384 /* Column and row sizing and page breaks.
385 sizing[TABLE_HORZ] is for columns, sizing[TABLE_VERT] is for rows. */
386 struct pivot_table_sizing sizing[TABLE_N_AXES];
388 /* Print settings. */
389 bool print_all_layers;
390 bool paginate_layers;
391 bool shrink_to_fit[TABLE_N_AXES];
392 bool top_continuation, bottom_continuation;
394 size_t n_orphan_lines;
396 /* Format settings. */
398 char decimal; /* Usually ',' or '.'. */
399 char grouping; /* Usually '.' or ','. */
400 char *ccs[5]; /* Custom currency. */
403 /* Command information. */
404 char *command_local; /* May be NULL. */
405 char *command_c; /* May be NULL. */
406 char *language; /* May be NULL. */
407 char *locale; /* May be NULL. */
409 /* Source information. */
410 char *dataset; /* May be NULL. */
411 char *datafile; /* May be NULL. */
412 time_t date; /* May be 0 if unknown. */
415 struct pivot_footnote **footnotes;
416 size_t n_footnotes, allocated_footnotes;
419 struct pivot_value *title;
420 struct pivot_value *subtype; /* Same as pivot_item's subtype. */
421 struct pivot_value *corner_text;
422 struct pivot_value *caption;
426 struct area_style areas[PIVOT_N_AREAS];
427 struct table_border_style borders[PIVOT_N_BORDERS];
430 struct pivot_dimension **dimensions;
433 /* Allocation of dimensions to rows, columns, and layers. */
434 struct pivot_axis axes[PIVOT_N_AXES];
436 struct hmap cells; /* Contains "struct pivot_cell"s. */
439 /* Creating and destroy pivot tables. */
440 struct pivot_table *pivot_table_create (const char *title);
441 struct pivot_table *pivot_table_create__ (struct pivot_value *title);
442 struct pivot_table *pivot_table_create_for_text (struct pivot_value *title,
443 struct pivot_value *content);
444 void pivot_table_destroy (struct pivot_table *);
446 /* Format of PIVOT_RC_COUNT cells. */
447 void pivot_table_set_weight_var (struct pivot_table *,
448 const struct variable *);
449 void pivot_table_set_weight_format (struct pivot_table *,
450 const struct fmt_spec *);
453 bool pivot_table_is_empty (const struct pivot_table *);
456 void pivot_table_submit (struct pivot_table *);
459 void pivot_table_put (struct pivot_table *, const size_t *dindexes, size_t n,
460 struct pivot_value *);
461 void pivot_table_put1 (struct pivot_table *, size_t idx1,
462 struct pivot_value *);
463 void pivot_table_put2 (struct pivot_table *, size_t idx1, size_t idx2,
464 struct pivot_value *);
465 void pivot_table_put3 (struct pivot_table *, size_t idx1, size_t idx2,
466 size_t idx3, struct pivot_value *);
467 void pivot_table_put4 (struct pivot_table *, size_t idx1, size_t idx2,
468 size_t idx3, size_t idx4, struct pivot_value *);
470 const struct pivot_value *pivot_table_get (const struct pivot_table *,
471 const size_t *dindexes);
473 struct pivot_value *pivot_table_get_rw (struct pivot_table *,
474 const size_t *dindexes);
478 Use pivot_table_create_footnote() to create a footnote.
479 Use pivot_value_add_footnote() to add a reference to a footnote. */
480 struct pivot_footnote
483 struct pivot_value *content;
484 struct pivot_value *marker;
487 struct pivot_footnote *pivot_table_create_footnote (
488 struct pivot_table *, struct pivot_value *content);
489 struct pivot_footnote *pivot_table_create_footnote__ (
490 struct pivot_table *, size_t idx,
491 struct pivot_value *marker, struct pivot_value *content);
493 void pivot_footnote_destroy (struct pivot_footnote *);
496 void pivot_table_convert_indexes_ptod (const struct pivot_table *,
497 const size_t *pindexes[PIVOT_N_AXES],
499 size_t *pivot_table_enumerate_axis (const struct pivot_table *,
500 enum pivot_axis_type,
501 const size_t *layer_indexes,
502 bool omit_empty, size_t *n);
503 #define PIVOT_ENUMERATION_FOR_EACH(INDEXES, ENUMERATION, AXIS) \
504 for ((INDEXES) = (ENUMERATION); *(INDEXES) != SIZE_MAX; \
505 (INDEXES) += MAX (1, (AXIS)->n_dimensions))
507 void pivot_table_assign_label_depth (struct pivot_table *);
509 void pivot_table_dump (const struct pivot_table *, int indentation);
513 enum pivot_value_type
515 PIVOT_VALUE_NUMERIC, /* A value of a numeric variable. */
516 PIVOT_VALUE_STRING, /* A value of a string variable. */
517 PIVOT_VALUE_VARIABLE, /* Name of a variable. */
518 PIVOT_VALUE_TEXT, /* Text. */
519 PIVOT_VALUE_TEMPLATE, /* Templated text. */
522 /* A pivot_value is the content of a single pivot table cell. A pivot_value is
523 also a pivot table's title, caption, footnote marker and contents, and so
526 A given pivot_value is one of:
528 1. A number resulting from a calculation (PIVOT_VALUE_NUMERIC). Use
529 pivot_value_new_number() to create such a pivot_value.
531 A numeric pivot_value has an associated display format (usually an F or
532 PCT format). This format can be set directly on the pivot_value, but
533 that is not usually the easiest way. Instead, it is usually true that
534 all of the values in a single category should have the same format
535 (e.g. all "Significance" values might use format F40.3), so PSPP makes
536 it easy to set the default format for a category while creating the
537 category. See pivot_dimension_create() for more details.
539 For numbers that should be displayed as integers,
540 pivot_value_new_integer() can occasionally be a useful special case.
542 2. A numeric or string value obtained from data (PIVOT_VALUE_NUMERIC or
543 PIVOT_VALUE_STRING). If such a value corresponds to a variable, then the
544 variable's name can be attached to the pivot_value. If the value has a
545 value label, then that can also be attached. When a label is present,
546 the user can control whether to show the value or the label or both.
548 Use pivot_value_new_var_value() to create pivot_values of these kinds.
550 3. A variable name (PIVOT_VALUE_VARIABLE). The variable label, if any, can
551 be attached too, and again the user can control whether to show the value
552 or the label or both.
554 4. A text string (PIVOT_VALUE_TEXT). The value stores the string in English
555 and translated into the output language (localized). Use
556 pivot_value_new_text() or pivot_value_new_text_format() for those cases.
557 In some cases, only an English or a localized version is available for
558 one reason or another, although this is regrettable; in those cases, use
559 pivot_value_new_user_text() or pivot_value_new_user_text_nocopy().
561 (There is also a PIVOT_VALUE_TEMPLATE but PSPP does not yet create these
568 A pivot_value may reference any number of footnotes. Use
569 pivot_value_add_footnote() to add a footnote reference. The footnotes being
570 referenced must first be created with pivot_table_create_footnote().
576 A pivot_value can have specific font and cell styles. Only the user should
581 struct font_style *font_style;
582 struct cell_style *cell_style;
586 struct pivot_footnote **footnotes;
589 enum pivot_value_type type;
592 /* PIVOT_VALUE_NUMERIC. */
595 double x; /* The numeric value. */
596 struct fmt_spec format; /* Format to display 'x'. */
597 char *var_name; /* May be NULL. */
598 char *value_label; /* May be NULL. */
599 enum settings_value_show show; /* Show value or label or both? */
603 /* PIVOT_VALUE_STRING. */
606 char *s; /* The string value. */
607 bool hex; /* Display in hex? */
608 char *var_name; /* May be NULL. */
609 char *value_label; /* May be NULL. */
610 enum settings_value_show show; /* Show value or label or both? */
614 /* PIVOT_VALUE_VARIABLE. */
618 char *var_label; /* May be NULL. */
619 enum settings_value_show show; /* Show name or label or both? */
623 /* PIVOT_VALUE_TEXT. */
626 char *local; /* Localized. */
627 char *c; /* English. */
628 char *id; /* Identifier. */
633 /* PIVOT_VALUE_TEMPLATE. */
637 struct pivot_argument *args;
644 /* Numbers resulting from calculations. */
645 struct pivot_value *pivot_value_new_number (double);
646 struct pivot_value *pivot_value_new_integer (double);
648 /* Values from data. */
649 struct pivot_value *pivot_value_new_var_value (
650 const struct variable *, const union value *);
651 struct pivot_value *pivot_value_new_value (const union value *, int width,
652 const struct fmt_spec *,
653 const char *encoding);
655 /* Values from variable names. */
656 struct pivot_value *pivot_value_new_variable (const struct variable *);
658 /* Values from text strings. */
659 struct pivot_value *pivot_value_new_text (const char *);
660 struct pivot_value *pivot_value_new_text_format (const char *, ...)
661 PRINTF_FORMAT (1, 2);
663 struct pivot_value *pivot_value_new_user_text (const char *, size_t length);
664 struct pivot_value *pivot_value_new_user_text_nocopy (char *);
667 void pivot_value_add_footnote (struct pivot_value *, struct pivot_footnote *);
669 /* Numeric formats. */
670 void pivot_value_set_rc (struct pivot_table *, struct pivot_value *,
673 /* Converting a pivot_value to a string for display. */
674 char *pivot_value_to_string (const struct pivot_value *,
675 enum settings_value_show show_values,
676 enum settings_value_show show_variables);
677 void pivot_value_format (const struct pivot_value *,
678 enum settings_value_show show_values,
679 enum settings_value_show show_variables,
681 bool pivot_value_format_body (const struct pivot_value *,
682 enum settings_value_show show_values,
683 enum settings_value_show show_variables,
686 void pivot_value_destroy (struct pivot_value *);
689 void pivot_value_get_style (struct pivot_value *,
690 const struct area_style *default_style,
691 struct area_style *);
692 void pivot_value_set_style (struct pivot_value *, const struct area_style *);
694 /* Template arguments. */
695 struct pivot_argument
698 struct pivot_value **values;
701 void pivot_argument_uninit (struct pivot_argument *);
703 #endif /* output/pivot-table.h */