UnionArray.h
Go to the documentation of this file.
1 // BSD 3-Clause License; see https://github.com/scikit-hep/awkward-1.0/blob/main/LICENSE
2 
3 #ifndef AWKWARD_UNIONARRAY_H_
4 #define AWKWARD_UNIONARRAY_H_
5 
6 #include <string>
7 #include <memory>
8 #include <vector>
9 
10 #include "awkward/common.h"
11 #include "awkward/Slice.h"
12 #include "awkward/Index.h"
13 #include "awkward/Content.h"
14 
15 namespace awkward {
19  class LIBAWKWARD_EXPORT_SYMBOL UnionForm: public Form {
20  public:
23  UnionForm(bool has_identities,
24  const util::Parameters& parameters,
25  const FormKey& form_key,
26  Index::Form tags,
27  Index::Form index,
28  const std::vector<FormPtr>& contents);
29 
30  Index::Form
31  tags() const;
32 
33  Index::Form
34  index() const;
35 
36  const std::vector<FormPtr>
37  contents() const;
38 
39  int64_t
40  numcontents() const;
41 
42  const FormPtr
43  content(int64_t index) const;
44 
45  const TypePtr
46  type(const util::TypeStrs& typestrs) const override;
47 
48  void
49  tojson_part(ToJson& builder, bool verbose) const override;
50 
51  const FormPtr
52  shallow_copy() const override;
53 
54  const FormPtr
55  with_form_key(const FormKey& form_key) const override;
56 
57  const std::string
58  purelist_parameter(const std::string& key) const override;
59 
60  bool
61  purelist_isregular() const override;
62 
63  int64_t
64  purelist_depth() const override;
65 
66  bool
67  dimension_optiontype() const override;
68 
69  const std::pair<int64_t, int64_t>
70  minmax_depth() const override;
71 
72  const std::pair<bool, int64_t>
73  branch_depth() const override;
74 
75  int64_t
76  numfields() const override;
77 
78  int64_t
79  fieldindex(const std::string& key) const override;
80 
81  const std::string
82  key(int64_t fieldindex) const override;
83 
84  bool
85  haskey(const std::string& key) const override;
86 
87  const std::vector<std::string>
88  keys() const override;
89 
90  bool
91  istuple() const override;
92 
93  bool
94  equal(const FormPtr& other,
95  bool check_identities,
96  bool check_parameters,
97  bool check_form_key,
98  bool compatibility_check) const override;
99 
100  const FormPtr
101  getitem_field(const std::string& key) const override;
102 
103  const FormPtr
104  getitem_fields(const std::vector<std::string>& keys) const override;
105 
106  private:
107  Index::Form tags_;
108  Index::Form index_;
109  const std::vector<FormPtr> contents_;
110  };
111 
123  template <typename T, typename I>
124  class
125 #ifdef AWKWARD_UNIONARRAY_NO_EXTERN_TEMPLATE
126  LIBAWKWARD_EXPORT_SYMBOL
127 #endif
128  UnionArrayOf: public Content {
129  public:
131  static const IndexOf<I>
132  sparse_index(int64_t len);
133 
135  static const IndexOf<I>
136  regular_index(const IndexOf<T>& tags);
137 
139  static const std::pair<IndexOf<T>, IndexOf<I>>
140  nested_tags_index(const Index64& offsets,
141  const std::vector<Index64>& counts);
142 
157  UnionArrayOf<T, I>(const IdentitiesPtr& identities,
158  const util::Parameters& parameters,
159  const IndexOf<T> tags,
160  const IndexOf<I>& index,
161  const ContentPtrVec& contents);
162 
169  const IndexOf<T>
170  tags() const;
171 
175  const IndexOf<I>
176  index() const;
177 
180  const ContentPtrVec
181  contents() const;
182 
184  int64_t
185  numcontents() const;
186 
188  const ContentPtr
189  content(int64_t index) const;
190 
195  const ContentPtr
196  project(int64_t index) const;
197 
210  const ContentPtr
211  simplify_uniontype(bool merge, bool mergebool) const;
212 
215  const std::string
216  classname() const override;
217 
218  void
219  setidentities() override;
220 
221  void
222  setidentities(const IdentitiesPtr& identities) override;
223 
224  const TypePtr
225  type(const util::TypeStrs& typestrs) const override;
226 
227  const FormPtr
228  form(bool materialize) const override;
229 
230  kernel::lib
231  kernels() const override;
232 
233  void
234  caches(std::vector<ArrayCachePtr>& out) const override;
235 
236  const std::string
237  tostring_part(const std::string& indent,
238  const std::string& pre,
239  const std::string& post) const override;
240 
241  void
242  tojson_part(ToJson& builder, bool include_beginendlist) const override;
243 
244  void
245  nbytes_part(std::map<size_t, int64_t>& largest) const override;
246 
250  int64_t
251  length() const override;
252 
253  const ContentPtr
254  shallow_copy() const override;
255 
256  const ContentPtr
257  deep_copy(bool copyarrays,
258  bool copyindexes,
259  bool copyidentities) const override;
260 
261  void
262  check_for_iteration() const override;
263 
264  const ContentPtr
265  getitem_nothing() const override;
266 
267  const ContentPtr
268  getitem_at(int64_t at) const override;
269 
270  const ContentPtr
271  getitem_at_nowrap(int64_t at) const override;
272 
273  const ContentPtr
274  getitem_range(int64_t start, int64_t stop) const override;
275 
276  const ContentPtr
277  getitem_range_nowrap(int64_t start, int64_t stop) const override;
278 
279  const ContentPtr
280  getitem_field(const std::string& key) const override;
281 
282  const ContentPtr
283  getitem_field(const std::string& key,
284  const Slice& only_fields) const override;
285 
286  const ContentPtr
287  getitem_fields(const std::vector<std::string>& keys) const override;
288 
289  const ContentPtr
290  getitem_fields(const std::vector<std::string>& keys,
291  const Slice& only_fields) const override;
292 
293  const ContentPtr
294  getitem_next(const SliceItemPtr& head,
295  const Slice& tail,
296  const Index64& advanced) const override;
297 
298  const ContentPtr
299  carry(const Index64& carry, bool allow_lazy) const override;
300 
301  int64_t
302  purelist_depth() const override;
303 
304  const std::pair<int64_t, int64_t>
305  minmax_depth() const override;
306 
307  const std::pair<bool, int64_t>
308  branch_depth() const override;
309 
310  int64_t
311  numfields() const override;
312 
313  int64_t
314  fieldindex(const std::string& key) const override;
315 
316  const std::string
317  key(int64_t fieldindex) const override;
318 
319  bool
320  haskey(const std::string& key) const override;
321 
322  const std::vector<std::string>
323  keys() const override;
324 
325  bool
326  istuple() const override;
327 
328  // operations
329  const std::string
330  validityerror(const std::string& path) const override;
331 
336  const ContentPtr
337  shallow_simplify() const override;
338 
339  const ContentPtr
340  num(int64_t axis, int64_t depth) const override;
341 
342  const std::pair<Index64, ContentPtr>
343  offsets_and_flattened(int64_t axis, int64_t depth) const override;
344 
345  bool
346  mergeable(const ContentPtr& other, bool mergebool) const override;
347 
348  bool
349  referentially_equal(const ContentPtr& other) const override;
350 
351  const ContentPtr
352  reverse_merge(const ContentPtr& other) const override;
353 
354  const std::pair<ContentPtrVec, ContentPtrVec>
355  merging_strategy(const ContentPtrVec& others) const override;
356 
357  const ContentPtr
358  mergemany(const ContentPtrVec& others) const override;
359 
360  const SliceItemPtr
361  asslice() const override;
362 
363  const ContentPtr
364  fillna(const ContentPtr& value) const override;
365 
366  const ContentPtr
367  rpad(int64_t target, int64_t axis, int64_t depth) const override;
368 
369  const ContentPtr
370  rpad_and_clip(int64_t target,
371  int64_t axis,
372  int64_t depth) const override;
373 
374  const ContentPtr
375  reduce_next(const Reducer& reducer,
376  int64_t negaxis,
377  const Index64& starts,
378  const Index64& shifts,
379  const Index64& parents,
380  int64_t outlength,
381  bool mask,
382  bool keepdims) const override;
383 
384  const ContentPtr
385  sort_next(int64_t negaxis,
386  const Index64& starts,
387  const Index64& parents,
388  int64_t outlength,
389  bool ascending,
390  bool stable) const override;
391 
392  const ContentPtr
393  argsort_next(int64_t negaxis,
394  const Index64& starts,
395  const Index64& shifts,
396  const Index64& parents,
397  int64_t outlength,
398  bool ascending,
399  bool stable) const override;
400 
401  const ContentPtr
402  localindex(int64_t axis, int64_t depth) const override;
403 
404  const ContentPtr
405  combinations(int64_t n,
406  bool replacement,
407  const util::RecordLookupPtr& recordlookup,
408  const util::Parameters& parameters,
409  int64_t axis,
410  int64_t depth) const override;
411 
412  const ContentPtr
413  getitem_next(const SliceAt& at,
414  const Slice& tail,
415  const Index64& advanced) const override;
416 
417  const ContentPtr
418  getitem_next(const SliceRange& range,
419  const Slice& tail,
420  const Index64& advanced) const override;
421 
422  const ContentPtr
423  getitem_next(const SliceArray64& array,
424  const Slice& tail,
425  const Index64& advanced) const override;
426 
427  const ContentPtr
428  getitem_next(const SliceJagged64& jagged,
429  const Slice& tail,
430  const Index64& advanced) const override;
431 
432  const ContentPtr
433  getitem_next_jagged(const Index64& slicestarts,
434  const Index64& slicestops,
435  const SliceArray64& slicecontent,
436  const Slice& tail) const override;
437 
438  const ContentPtr
439  getitem_next_jagged(const Index64& slicestarts,
440  const Index64& slicestops,
441  const SliceMissing64& slicecontent,
442  const Slice& tail) const override;
443 
444  const ContentPtr
445  getitem_next_jagged(const Index64& slicestarts,
446  const Index64& slicestops,
447  const SliceJagged64& slicecontent,
448  const Slice& tail) const override;
449 
450  const ContentPtr
451  copy_to(kernel::lib ptr_lib) const override;
452 
453  const ContentPtr
454  numbers_to_type(const std::string& name) const override;
455 
457  bool
458  is_unique() const override;
459 
461  const ContentPtr
462  unique() const override;
463 
465  bool
466  is_subrange_equal(const Index64& start, const Index64& stop) const override;
467 
468  protected:
469  template <typename S>
470  const ContentPtr
471  getitem_next_jagged_generic(const Index64& slicestarts,
472  const Index64& slicestops,
473  const S& slicecontent,
474  const Slice& tail) const;
475 
476  private:
477  const IndexOf<T> tags_;
478  const IndexOf<I> index_;
479  const ContentPtrVec contents_;
480  };
481 
482 #ifndef AWKWARD_UNIONARRAY_NO_EXTERN_TEMPLATE
483  extern template class UnionArrayOf<int8_t, int32_t>;
484  extern template class UnionArrayOf<int8_t, uint32_t>;
485  extern template class UnionArrayOf<int8_t, int64_t>;
486 #endif
487 
488  using UnionArray8_32 = UnionArrayOf<int8_t, int32_t>;
489  using UnionArray8_U32 = UnionArrayOf<int8_t, uint32_t>;
490  using UnionArray8_64 = UnionArrayOf<int8_t, int64_t>;
491 }
492 
493 #endif // AWKWARD_UNIONARRAY_H_
awkward::SliceRange
Represents a Python slice object (usual syntax: array[start:stop:step]).
Definition: Slice.h:93
LIBAWKWARD_EXPORT_SYMBOL
#define LIBAWKWARD_EXPORT_SYMBOL
Definition: common.h:45
awkward::util::key
const std::string key(const RecordLookupPtr &recordlookup, int64_t fieldindex, int64_t numfields)
Returns the key associated with a field index, given a RecordLookup and a number of fields...
awkward::Content
Abstract superclass of all array node types (flat hierarchy). Any Content can be nested within any ot...
Definition: Content.h:276
awkward::ToJson
Definition: json.h:19
awkward
Definition: BitMaskedArray.h:15
awkward::UnionForm
Form describing UnionArray.
Definition: UnionArray.h:19
len
int64_t len(const T &self)
Definition: content.h:46
awkward::util::name
std::string name(Mapping a, V value)
Definition: datetime_util.h:39
awkward::SliceMissingOf
Represents a SliceArrayOf, SliceMissingOf, or SliceJaggedOf with missing values: None (no equivalent ...
Definition: Slice.h:431
Content.h
awkward::Reducer
Abstract class for all reducer algorithms.
Definition: Reducer.h:20
awkward::kernel::combinations
ERROR combinations(kernel::lib ptr_lib, T *toindex, int64_t n, bool replacement, int64_t singlelen)
awkward::util::keys
const std::vector< std::string > keys(const RecordLookupPtr &recordlookup, int64_t numfields)
Returns a given RecordLookup as keys or generate anonymous ones form a number of fields.
awkward::Index::Form
Form
Integer type of an Index, used by ListForm, IndexedForm, etc.
Definition: Index.h:34
awkward::FormKey
std::shared_ptr< std::string > FormKey
Definition: Content.h:19
awkward::kernel::unique
ERROR unique(kernel::lib ptr_lib, T *toptr, int64_t length, int64_t *tolength)
awkward::SliceArrayOf
Represents an array of integers in a slice (possibly converted from an array of booleans).
Definition: Slice.h:225
awkward::util::Parameters
std::map< std::string, std::string > Parameters
Definition: util.h:165
awkward::kernel::copy_to
ERROR copy_to(kernel::lib to_lib, kernel::lib from_lib, void *to_ptr, void *from_ptr, int64_t bytelength)
Internal Function an array buffer from library FROM to library TO, usually between main memory and a ...
awkward::FormPtr
std::shared_ptr< Form > FormPtr
Definition: Content.h:18
awkward::util::fieldindex
int64_t fieldindex(const RecordLookupPtr &recordlookup, const std::string &key, int64_t numfields)
Returns the field index associated with a key, given a RecordLookup and a number of fields...
awkward::SliceJaggedOf
Represents an array of nested lists, where the content may be SliceArrayOf, SliceMissingOf, or SliceJaggedOf (no equivalent in NumPy).
Definition: Slice.h:511
awkward::ContentPtrVec
std::vector< std::shared_ptr< Content > > ContentPtrVec
Definition: Content.h:16
awkward::util::value
Mapping::value_type::value_type value(Mapping a, const std::string &name)
Definition: datetime_util.h:49
awkward::Slice
A sequence of SliceItem objects representing a tuple passed to Python&#39;s __getitem__.
Definition: Slice.h:585
awkward::util::haskey
bool haskey(const RecordLookupPtr &recordlookup, const std::string &key, int64_t numfields)
Returns true if a RecordLookup has a given key; false otherwise.
awkward::UnionArrayOf
Represents heterogeneous data by interleaving several contents, indicating which is relevant at a giv...
Definition: UnionArray.h:124
awkward::TypePtr
std::shared_ptr< Type > TypePtr
Definition: Content.h:23
awkward::util::RecordLookupPtr
std::shared_ptr< RecordLookup > RecordLookupPtr
Definition: util.h:130
awkward::SliceAt
Represents an integer in a tuple of slices passed to __getitem__ in Python.
Definition: Slice.h:58
awkward::util::TypeStrs
std::map< std::string, std::string > TypeStrs
Definition: util.h:215
awkward::Form
Abstract superclass of all array node forms, which expresses the nesting structure without any large ...
Definition: Content.h:39
awkward::IdentitiesPtr
std::shared_ptr< Identities > IdentitiesPtr
Definition: Identities.h:16
awkward::ContentPtr
std::shared_ptr< Content > ContentPtr
Definition: ArrayBuilder.h:16
awkward::SliceItemPtr
std::shared_ptr< SliceItem > SliceItemPtr
Definition: Slice.h:15
awkward::kernel::lib
lib
Definition: kernel-dispatch.h:20