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 /* Minimum and maximum column width, in 1/96" units. */
157 /* Specific column widths, in 1/96" units. */
161 /* Specific page breaks: 0-based columns after which a page break must
162 occur, e.g. a value of 1 requests a break after the second column. */
166 /* Keeps: columns to keep together on a page if possible. */
167 struct pivot_keep *keeps;
171 void pivot_table_sizing_uninit (struct pivot_table_sizing *);
173 /* A set of columns to keep together on a page if possible, e.g. ofs=1, n=10
174 requests keeping together the 2nd through 11th columns. */
177 size_t ofs; /* 0-based first column. */
178 size_t n; /* Number of columns. */
192 const char *pivot_axis_type_to_string (enum pivot_axis_type);
194 /* An axis within a pivot table. */
197 /* dimensions[0] is the innermost dimension,
198 dimensions[1] is the next outer dimension,
200 dimensions[n_dimensions - 1] is the outermost dimension. */
201 struct pivot_dimension **dimensions;
204 /* The number of rows or columns along the axis,
205 that is, the product of dimension[*]->n_leaves.
206 It is 0 if any dimension has 0 leaves. */
209 /* Sum of dimensions[*]->label_depth. */
213 /* Successively assigns to INDEXES (which should be a "size_t *") each of the
214 combinations of the categories in AXIS's dimensions, in lexicographic order
215 with the innermost dimension iterating most quickly.
217 The value assigned to INDEXES is dynamically allocated. If the client
218 breaks out of the loop prematurely, it needs to free it with free(). */
219 #define PIVOT_AXIS_FOR_EACH(INDEXES, AXIS) \
220 for ((INDEXES) = NULL; \
221 ((INDEXES) = pivot_axis_iterator_next (INDEXES, AXIS)) != NULL; )
222 size_t *pivot_axis_iterator_next (size_t *indexes, const struct pivot_axis *);
226 A pivot_dimension identifies the categories associated with a single
227 dimension within a multidimensional pivot table.
229 A dimension contains a collection of categories, which are the leaves in a
232 (A dimension or a group can contain zero categories, but this is unusual.
233 If a dimension contains no categories, then its table cannot contain any
236 struct pivot_dimension
238 /* table->axes[axis_type]->dimensions[level] == dimension. */
239 struct pivot_table *table;
240 enum pivot_axis_type axis_type;
241 size_t level; /* 0 for innermost dimension within axis. */
243 /* table->dimensions[top_index] == dimension. */
246 /* Hierarchy of categories within the dimension. The groups and categories
247 are sorted in the order that should be used for display. This might be
248 different from the original order produced for output if the user
251 The root must always be a group, although it is allowed to have no
253 struct pivot_category *root;
255 /* All of the leaves reachable via the root.
257 The indexing for presentation_leaves is presentation order, thus
258 presentation_leaves[i]->presentation_index == i. This order is the same
259 as would be produced by an in-order traversal of the groups. It is the
260 order into which the user reordered or sorted the categories.
262 The indexing for data_leaves is that used for idx[] in struct
263 pivot_cell, thus data_leaves[i]->data_index == i. This might differ
264 from what an in-order traversal of 'root' would yield, if the user
265 reordered categories. */
266 struct pivot_category **data_leaves;
267 struct pivot_category **presentation_leaves;
268 size_t n_leaves, allocated_leaves;
271 bool hide_all_labels;
273 /* Number of rows or columns needed to express the labels. */
277 struct pivot_dimension *pivot_dimension_create (
278 struct pivot_table *, enum pivot_axis_type, const char *name, ...)
280 #define pivot_dimension_create(...) \
281 pivot_dimension_create(__VA_ARGS__, NULL_SENTINEL)
282 struct pivot_dimension *pivot_dimension_create__ (struct pivot_table *,
283 enum pivot_axis_type,
284 struct pivot_value *name);
286 void pivot_dimension_destroy (struct pivot_dimension *);
288 void pivot_dimension_dump (const struct pivot_dimension *, int indentation);
290 /* A pivot_category is a leaf (a category) or a group:
292 - For a leaf, neither index is SIZE_MAX.
294 - For a group, both indexes are SIZE_MAX.
296 Do not use 'subs' or 'n_subs' to determine whether a category is a group,
297 because a group may (pathologically) have no leaves. */
298 struct pivot_category
300 struct pivot_value *name;
301 struct pivot_category *parent;
302 struct pivot_dimension *dimension;
303 size_t label_depth, extra_depth;
307 If show_label is true, then the group itself has a row (or a column)
308 giving the group's name. Otherwise, the group's own name is not
310 struct pivot_category **subs; /* Child categories or groups. */
311 size_t n_subs, allocated_subs;
312 bool show_label; /* Display a label for the group itself? */
313 bool show_label_in_corner;
316 struct fmt_spec format;
317 size_t group_index; /* In ->parent->subs[]. */
318 size_t data_index; /* In ->dimension->data_leaves[]. */
319 size_t presentation_index; /* In ->dimension->presentation_leaves[]. */
323 pivot_category_is_group (const struct pivot_category *category)
325 return category->data_index == SIZE_MAX;
329 pivot_category_is_leaf (const struct pivot_category *category)
331 return !pivot_category_is_group (category);
334 /* Creating leaf categories. */
335 int pivot_category_create_leaves (struct pivot_category *parent, ...)
337 #define pivot_category_create_leaves(...) \
338 pivot_category_create_leaves(__VA_ARGS__, NULL_SENTINEL)
340 int pivot_category_create_leaf (
341 struct pivot_category *parent, struct pivot_value *name);
342 int pivot_category_create_leaf_rc (
343 struct pivot_category *parent, struct pivot_value *name, const char *rc);
345 /* Creating category groups. */
346 struct pivot_category *pivot_category_create_group (
347 struct pivot_category *parent, const char *name, ...) SENTINEL (0);
348 #define pivot_category_create_group(...) \
349 pivot_category_create_group(__VA_ARGS__, NULL_SENTINEL)
350 struct pivot_category *pivot_category_create_group__ (
351 struct pivot_category *parent, struct pivot_value *name);
353 void pivot_category_destroy (struct pivot_category *);
355 /* Pivot result classes.
357 These are used to mark leaf categories as having particular types of data,
358 to set their numeric formats. The formats that actually get used for these
359 classes are in the result_classes[] global array in pivot-table.c, except
360 that PIVOT_RC_OTHER comes from settings_get_format() and PIVOT_RC_COUNT
361 should come from the weight variable in the dataset's dictionary. */
362 #define PIVOT_RC_OTHER ("RC_OTHER")
363 #define PIVOT_RC_INTEGER ("RC_INTEGER")
364 #define PIVOT_RC_CORRELATION ("RC_CORRELATIONS")
365 #define PIVOT_RC_SIGNIFICANCE ("RC_SIGNIFICANCE")
366 #define PIVOT_RC_PERCENT ("RC_PERCENT")
367 #define PIVOT_RC_RESIDUAL ("RC_RESIDUAL")
368 #define PIVOT_RC_COUNT ("RC_COUNT")
370 bool pivot_result_class_change (const char *, const struct fmt_spec *);
372 /* A pivot table. See the top of this file for more information. */
375 /* Reference count. A pivot_table may be shared between multiple owners,
376 indicated by a reference count greater than 1. When this is the case,
377 the pivot_table must not be modified. */
380 /* Display settings. */
381 bool rotate_inner_column_labels;
382 bool rotate_outer_row_labels;
383 bool row_labels_in_corner;
384 bool show_grid_lines;
385 bool omit_empty; /* Omit empty rows and columns? */
386 size_t *current_layer; /* axis[PIVOT_AXIS_LAYER].n_dimensions elements. */
388 enum settings_value_show show_values;
389 enum settings_value_show show_variables;
390 struct fmt_spec weight_format;
392 /* Footnote display settings. */
393 bool show_numeric_markers;
394 bool footnote_marker_superscripts;
396 /* Column and row sizing and page breaks.
397 sizing[TABLE_HORZ] is for columns, sizing[TABLE_VERT] is for rows. */
398 struct pivot_table_sizing sizing[TABLE_N_AXES];
400 /* Print settings. */
401 bool print_all_layers;
402 bool paginate_layers;
403 bool shrink_to_fit[TABLE_N_AXES];
404 bool top_continuation, bottom_continuation;
406 size_t n_orphan_lines;
408 /* Format settings. */
410 char decimal; /* Usually ',' or '.'. */
411 char grouping; /* Usually '.' or ','. */
412 char *ccs[5]; /* Custom currency. */
415 /* Command information. */
416 char *command_local; /* May be NULL. */
417 char *command_c; /* May be NULL. */
418 char *language; /* May be NULL. */
419 char *locale; /* May be NULL. */
421 /* Source information. */
422 char *dataset; /* May be NULL. */
423 char *datafile; /* May be NULL. */
424 time_t date; /* May be 0 if unknown. */
427 struct pivot_footnote **footnotes;
428 size_t n_footnotes, allocated_footnotes;
431 struct pivot_value *title;
432 struct pivot_value *subtype; /* Same as pivot_item's subtype. */
433 struct pivot_value *corner_text;
434 struct pivot_value *caption;
438 struct area_style areas[PIVOT_N_AREAS];
439 struct table_border_style borders[PIVOT_N_BORDERS];
442 struct pivot_dimension **dimensions;
445 /* Allocation of dimensions to rows, columns, and layers. */
446 struct pivot_axis axes[PIVOT_N_AXES];
448 struct hmap cells; /* Contains "struct pivot_cell"s. */
451 /* Creating and destroy pivot tables. */
452 struct pivot_table *pivot_table_create (const char *title);
453 struct pivot_table *pivot_table_create__ (struct pivot_value *title);
454 struct pivot_table *pivot_table_create_for_text (struct pivot_value *title,
455 struct pivot_value *content);
457 struct pivot_table *pivot_table_ref (const struct pivot_table *);
458 void pivot_table_unref (struct pivot_table *);
459 bool pivot_table_is_shared (const struct pivot_table *);
461 /* Format of PIVOT_RC_COUNT cells. */
462 void pivot_table_set_weight_var (struct pivot_table *,
463 const struct variable *);
464 void pivot_table_set_weight_format (struct pivot_table *,
465 const struct fmt_spec *);
468 bool pivot_table_is_empty (const struct pivot_table *);
471 void pivot_table_submit (struct pivot_table *);
474 void pivot_table_put (struct pivot_table *, const size_t *dindexes, size_t n,
475 struct pivot_value *);
476 void pivot_table_put1 (struct pivot_table *, size_t idx1,
477 struct pivot_value *);
478 void pivot_table_put2 (struct pivot_table *, size_t idx1, size_t idx2,
479 struct pivot_value *);
480 void pivot_table_put3 (struct pivot_table *, size_t idx1, size_t idx2,
481 size_t idx3, struct pivot_value *);
482 void pivot_table_put4 (struct pivot_table *, size_t idx1, size_t idx2,
483 size_t idx3, size_t idx4, struct pivot_value *);
485 const struct pivot_value *pivot_table_get (const struct pivot_table *,
486 const size_t *dindexes);
488 struct pivot_value *pivot_table_get_rw (struct pivot_table *,
489 const size_t *dindexes);
493 Use pivot_table_create_footnote() to create a footnote.
494 Use pivot_value_add_footnote() to add a reference to a footnote. */
495 struct pivot_footnote
498 struct pivot_value *content;
499 struct pivot_value *marker;
502 struct pivot_footnote *pivot_table_create_footnote (
503 struct pivot_table *, struct pivot_value *content);
504 struct pivot_footnote *pivot_table_create_footnote__ (
505 struct pivot_table *, size_t idx,
506 struct pivot_value *marker, struct pivot_value *content);
508 void pivot_footnote_destroy (struct pivot_footnote *);
511 void pivot_table_convert_indexes_ptod (const struct pivot_table *,
512 const size_t *pindexes[PIVOT_N_AXES],
514 size_t *pivot_table_enumerate_axis (const struct pivot_table *,
515 enum pivot_axis_type,
516 const size_t *layer_indexes,
517 bool omit_empty, size_t *n);
518 #define PIVOT_ENUMERATION_FOR_EACH(INDEXES, ENUMERATION, AXIS) \
519 for ((INDEXES) = (ENUMERATION); *(INDEXES) != SIZE_MAX; \
520 (INDEXES) += MAX (1, (AXIS)->n_dimensions))
522 void pivot_table_assign_label_depth (struct pivot_table *);
524 void pivot_table_dump (const struct pivot_table *, int indentation);
528 enum pivot_value_type
530 PIVOT_VALUE_NUMERIC, /* A value of a numeric variable. */
531 PIVOT_VALUE_STRING, /* A value of a string variable. */
532 PIVOT_VALUE_VARIABLE, /* Name of a variable. */
533 PIVOT_VALUE_TEXT, /* Text. */
534 PIVOT_VALUE_TEMPLATE, /* Templated text. */
537 /* A pivot_value is the content of a single pivot table cell. A pivot_value is
538 also a pivot table's title, caption, footnote marker and contents, and so
541 A given pivot_value is one of:
543 1. A number resulting from a calculation (PIVOT_VALUE_NUMERIC). Use
544 pivot_value_new_number() to create such a pivot_value.
546 A numeric pivot_value has an associated display format (usually an F or
547 PCT format). This format can be set directly on the pivot_value, but
548 that is not usually the easiest way. Instead, it is usually true that
549 all of the values in a single category should have the same format
550 (e.g. all "Significance" values might use format F40.3), so PSPP makes
551 it easy to set the default format for a category while creating the
552 category. See pivot_dimension_create() for more details.
554 For numbers that should be displayed as integers,
555 pivot_value_new_integer() can occasionally be a useful special case.
557 2. A numeric or string value obtained from data (PIVOT_VALUE_NUMERIC or
558 PIVOT_VALUE_STRING). If such a value corresponds to a variable, then the
559 variable's name can be attached to the pivot_value. If the value has a
560 value label, then that can also be attached. When a label is present,
561 the user can control whether to show the value or the label or both.
563 Use pivot_value_new_var_value() to create pivot_values of these kinds.
565 3. A variable name (PIVOT_VALUE_VARIABLE). The variable label, if any, can
566 be attached too, and again the user can control whether to show the value
567 or the label or both.
569 4. A text string (PIVOT_VALUE_TEXT). The value stores the string in English
570 and translated into the output language (localized). Use
571 pivot_value_new_text() or pivot_value_new_text_format() for those cases.
572 In some cases, only an English or a localized version is available for
573 one reason or another, although this is regrettable; in those cases, use
574 pivot_value_new_user_text() or pivot_value_new_user_text_nocopy().
576 (There is also a PIVOT_VALUE_TEMPLATE but PSPP does not yet create these
583 A pivot_value may reference any number of footnotes. Use
584 pivot_value_add_footnote() to add a footnote reference. The footnotes being
585 referenced must first be created with pivot_table_create_footnote().
591 A pivot_value can have specific font and cell styles. Only the user should
596 struct font_style *font_style;
597 struct cell_style *cell_style;
601 const struct pivot_footnote **footnotes;
604 enum pivot_value_type type;
607 /* PIVOT_VALUE_NUMERIC. */
610 double x; /* The numeric value. */
611 struct fmt_spec format; /* Format to display 'x'. */
612 char *var_name; /* May be NULL. */
613 char *value_label; /* May be NULL. */
614 enum settings_value_show show; /* Show value or label or both? */
618 /* PIVOT_VALUE_STRING. */
621 char *s; /* The string value. */
622 bool hex; /* Display in hex? */
623 char *var_name; /* May be NULL. */
624 char *value_label; /* May be NULL. */
625 enum settings_value_show show; /* Show value or label or both? */
629 /* PIVOT_VALUE_VARIABLE. */
633 char *var_label; /* May be NULL. */
634 enum settings_value_show show; /* Show name or label or both? */
638 /* PIVOT_VALUE_TEXT. */
641 char *local; /* Localized. */
642 char *c; /* English. */
643 char *id; /* Identifier. */
648 /* PIVOT_VALUE_TEMPLATE. */
652 struct pivot_argument *args;
659 /* Numbers resulting from calculations. */
660 struct pivot_value *pivot_value_new_number (double);
661 struct pivot_value *pivot_value_new_integer (double);
663 /* Values from data. */
664 struct pivot_value *pivot_value_new_var_value (
665 const struct variable *, const union value *);
666 struct pivot_value *pivot_value_new_value (const union value *, int width,
667 const struct fmt_spec *,
668 const char *encoding);
670 /* Values from variable names. */
671 struct pivot_value *pivot_value_new_variable (const struct variable *);
673 /* Values from text strings. */
674 struct pivot_value *pivot_value_new_text (const char *);
675 struct pivot_value *pivot_value_new_text_format (const char *, ...)
676 PRINTF_FORMAT (1, 2);
678 struct pivot_value *pivot_value_new_user_text (const char *, size_t length);
679 struct pivot_value *pivot_value_new_user_text_nocopy (char *);
682 void pivot_value_add_footnote (struct pivot_value *, const struct pivot_footnote *);
684 /* Numeric formats. */
685 void pivot_value_set_rc (const struct pivot_table *, struct pivot_value *,
688 /* Converting a pivot_value to a string for display. */
689 char *pivot_value_to_string (const struct pivot_value *,
690 enum settings_value_show show_values,
691 enum settings_value_show show_variables);
692 void pivot_value_format (const struct pivot_value *,
693 enum settings_value_show show_values,
694 enum settings_value_show show_variables,
696 bool pivot_value_format_body (const struct pivot_value *,
697 enum settings_value_show show_values,
698 enum settings_value_show show_variables,
701 void pivot_value_destroy (struct pivot_value *);
704 void pivot_value_get_style (struct pivot_value *,
705 const struct area_style *default_style,
706 struct area_style *);
707 void pivot_value_set_style (struct pivot_value *, const struct area_style *);
709 /* Template arguments. */
710 struct pivot_argument
713 struct pivot_value **values;
716 void pivot_argument_uninit (struct pivot_argument *);
718 #endif /* output/pivot-table.h */