|
1 """ |
|
2 Query subclasses which provide extra functionality beyond simple data retrieval. |
|
3 """ |
|
4 |
|
5 from django.core.exceptions import FieldError |
|
6 from django.db.models.sql.constants import * |
|
7 from django.db.models.sql.datastructures import Date |
|
8 from django.db.models.sql.expressions import SQLEvaluator |
|
9 from django.db.models.sql.query import Query |
|
10 from django.db.models.sql.where import AND, Constraint |
|
11 |
|
12 __all__ = ['DeleteQuery', 'UpdateQuery', 'InsertQuery', 'DateQuery', |
|
13 'AggregateQuery'] |
|
14 |
|
15 class DeleteQuery(Query): |
|
16 """ |
|
17 Delete queries are done through this class, since they are more constrained |
|
18 than general queries. |
|
19 """ |
|
20 def as_sql(self): |
|
21 """ |
|
22 Creates the SQL for this query. Returns the SQL string and list of |
|
23 parameters. |
|
24 """ |
|
25 assert len(self.tables) == 1, \ |
|
26 "Can only delete from one table at a time." |
|
27 result = ['DELETE FROM %s' % self.quote_name_unless_alias(self.tables[0])] |
|
28 where, params = self.where.as_sql() |
|
29 result.append('WHERE %s' % where) |
|
30 return ' '.join(result), tuple(params) |
|
31 |
|
32 def do_query(self, table, where): |
|
33 self.tables = [table] |
|
34 self.where = where |
|
35 self.execute_sql(None) |
|
36 |
|
37 def delete_batch_related(self, pk_list): |
|
38 """ |
|
39 Set up and execute delete queries for all the objects related to the |
|
40 primary key values in pk_list. To delete the objects themselves, use |
|
41 the delete_batch() method. |
|
42 |
|
43 More than one physical query may be executed if there are a |
|
44 lot of values in pk_list. |
|
45 """ |
|
46 from django.contrib.contenttypes import generic |
|
47 cls = self.model |
|
48 for related in cls._meta.get_all_related_many_to_many_objects(): |
|
49 if not isinstance(related.field, generic.GenericRelation): |
|
50 for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE): |
|
51 where = self.where_class() |
|
52 where.add((Constraint(None, |
|
53 related.field.m2m_reverse_name(), related.field), |
|
54 'in', |
|
55 pk_list[offset : offset+GET_ITERATOR_CHUNK_SIZE]), |
|
56 AND) |
|
57 self.do_query(related.field.m2m_db_table(), where) |
|
58 |
|
59 for f in cls._meta.many_to_many: |
|
60 w1 = self.where_class() |
|
61 if isinstance(f, generic.GenericRelation): |
|
62 from django.contrib.contenttypes.models import ContentType |
|
63 field = f.rel.to._meta.get_field(f.content_type_field_name) |
|
64 w1.add((Constraint(None, field.column, field), 'exact', |
|
65 ContentType.objects.get_for_model(cls).id), AND) |
|
66 for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE): |
|
67 where = self.where_class() |
|
68 where.add((Constraint(None, f.m2m_column_name(), f), 'in', |
|
69 pk_list[offset : offset + GET_ITERATOR_CHUNK_SIZE]), |
|
70 AND) |
|
71 if w1: |
|
72 where.add(w1, AND) |
|
73 self.do_query(f.m2m_db_table(), where) |
|
74 |
|
75 def delete_batch(self, pk_list): |
|
76 """ |
|
77 Set up and execute delete queries for all the objects in pk_list. This |
|
78 should be called after delete_batch_related(), if necessary. |
|
79 |
|
80 More than one physical query may be executed if there are a |
|
81 lot of values in pk_list. |
|
82 """ |
|
83 for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE): |
|
84 where = self.where_class() |
|
85 field = self.model._meta.pk |
|
86 where.add((Constraint(None, field.column, field), 'in', |
|
87 pk_list[offset : offset + GET_ITERATOR_CHUNK_SIZE]), AND) |
|
88 self.do_query(self.model._meta.db_table, where) |
|
89 |
|
90 class UpdateQuery(Query): |
|
91 """ |
|
92 Represents an "update" SQL query. |
|
93 """ |
|
94 def __init__(self, *args, **kwargs): |
|
95 super(UpdateQuery, self).__init__(*args, **kwargs) |
|
96 self._setup_query() |
|
97 |
|
98 def _setup_query(self): |
|
99 """ |
|
100 Runs on initialization and after cloning. Any attributes that would |
|
101 normally be set in __init__ should go in here, instead, so that they |
|
102 are also set up after a clone() call. |
|
103 """ |
|
104 self.values = [] |
|
105 self.related_ids = None |
|
106 if not hasattr(self, 'related_updates'): |
|
107 self.related_updates = {} |
|
108 |
|
109 def clone(self, klass=None, **kwargs): |
|
110 return super(UpdateQuery, self).clone(klass, |
|
111 related_updates=self.related_updates.copy(), **kwargs) |
|
112 |
|
113 def execute_sql(self, result_type=None): |
|
114 """ |
|
115 Execute the specified update. Returns the number of rows affected by |
|
116 the primary update query. The "primary update query" is the first |
|
117 non-empty query that is executed. Row counts for any subsequent, |
|
118 related queries are not available. |
|
119 """ |
|
120 cursor = super(UpdateQuery, self).execute_sql(result_type) |
|
121 rows = cursor and cursor.rowcount or 0 |
|
122 is_empty = cursor is None |
|
123 del cursor |
|
124 for query in self.get_related_updates(): |
|
125 aux_rows = query.execute_sql(result_type) |
|
126 if is_empty: |
|
127 rows = aux_rows |
|
128 is_empty = False |
|
129 return rows |
|
130 |
|
131 def as_sql(self): |
|
132 """ |
|
133 Creates the SQL for this query. Returns the SQL string and list of |
|
134 parameters. |
|
135 """ |
|
136 self.pre_sql_setup() |
|
137 if not self.values: |
|
138 return '', () |
|
139 table = self.tables[0] |
|
140 qn = self.quote_name_unless_alias |
|
141 result = ['UPDATE %s' % qn(table)] |
|
142 result.append('SET') |
|
143 values, update_params = [], [] |
|
144 for name, val, placeholder in self.values: |
|
145 if hasattr(val, 'as_sql'): |
|
146 sql, params = val.as_sql(qn) |
|
147 values.append('%s = %s' % (qn(name), sql)) |
|
148 update_params.extend(params) |
|
149 elif val is not None: |
|
150 values.append('%s = %s' % (qn(name), placeholder)) |
|
151 update_params.append(val) |
|
152 else: |
|
153 values.append('%s = NULL' % qn(name)) |
|
154 result.append(', '.join(values)) |
|
155 where, params = self.where.as_sql() |
|
156 if where: |
|
157 result.append('WHERE %s' % where) |
|
158 return ' '.join(result), tuple(update_params + params) |
|
159 |
|
160 def pre_sql_setup(self): |
|
161 """ |
|
162 If the update depends on results from other tables, we need to do some |
|
163 munging of the "where" conditions to match the format required for |
|
164 (portable) SQL updates. That is done here. |
|
165 |
|
166 Further, if we are going to be running multiple updates, we pull out |
|
167 the id values to update at this point so that they don't change as a |
|
168 result of the progressive updates. |
|
169 """ |
|
170 self.select_related = False |
|
171 self.clear_ordering(True) |
|
172 super(UpdateQuery, self).pre_sql_setup() |
|
173 count = self.count_active_tables() |
|
174 if not self.related_updates and count == 1: |
|
175 return |
|
176 |
|
177 # We need to use a sub-select in the where clause to filter on things |
|
178 # from other tables. |
|
179 query = self.clone(klass=Query) |
|
180 query.bump_prefix() |
|
181 query.extra = {} |
|
182 query.select = [] |
|
183 query.add_fields([query.model._meta.pk.name]) |
|
184 must_pre_select = count > 1 and not self.connection.features.update_can_self_select |
|
185 |
|
186 # Now we adjust the current query: reset the where clause and get rid |
|
187 # of all the tables we don't need (since they're in the sub-select). |
|
188 self.where = self.where_class() |
|
189 if self.related_updates or must_pre_select: |
|
190 # Either we're using the idents in multiple update queries (so |
|
191 # don't want them to change), or the db backend doesn't support |
|
192 # selecting from the updating table (e.g. MySQL). |
|
193 idents = [] |
|
194 for rows in query.execute_sql(MULTI): |
|
195 idents.extend([r[0] for r in rows]) |
|
196 self.add_filter(('pk__in', idents)) |
|
197 self.related_ids = idents |
|
198 else: |
|
199 # The fast path. Filters and updates in one query. |
|
200 self.add_filter(('pk__in', query)) |
|
201 for alias in self.tables[1:]: |
|
202 self.alias_refcount[alias] = 0 |
|
203 |
|
204 def clear_related(self, related_field, pk_list): |
|
205 """ |
|
206 Set up and execute an update query that clears related entries for the |
|
207 keys in pk_list. |
|
208 |
|
209 This is used by the QuerySet.delete_objects() method. |
|
210 """ |
|
211 for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE): |
|
212 self.where = self.where_class() |
|
213 f = self.model._meta.pk |
|
214 self.where.add((Constraint(None, f.column, f), 'in', |
|
215 pk_list[offset : offset + GET_ITERATOR_CHUNK_SIZE]), |
|
216 AND) |
|
217 self.values = [(related_field.column, None, '%s')] |
|
218 self.execute_sql(None) |
|
219 |
|
220 def add_update_values(self, values): |
|
221 """ |
|
222 Convert a dictionary of field name to value mappings into an update |
|
223 query. This is the entry point for the public update() method on |
|
224 querysets. |
|
225 """ |
|
226 values_seq = [] |
|
227 for name, val in values.iteritems(): |
|
228 field, model, direct, m2m = self.model._meta.get_field_by_name(name) |
|
229 if not direct or m2m: |
|
230 raise FieldError('Cannot update model field %r (only non-relations and foreign keys permitted).' % field) |
|
231 values_seq.append((field, model, val)) |
|
232 return self.add_update_fields(values_seq) |
|
233 |
|
234 def add_update_fields(self, values_seq): |
|
235 """ |
|
236 Turn a sequence of (field, model, value) triples into an update query. |
|
237 Used by add_update_values() as well as the "fast" update path when |
|
238 saving models. |
|
239 """ |
|
240 from django.db.models.base import Model |
|
241 for field, model, val in values_seq: |
|
242 if hasattr(val, 'prepare_database_save'): |
|
243 val = val.prepare_database_save(field) |
|
244 else: |
|
245 val = field.get_db_prep_save(val) |
|
246 |
|
247 # Getting the placeholder for the field. |
|
248 if hasattr(field, 'get_placeholder'): |
|
249 placeholder = field.get_placeholder(val) |
|
250 else: |
|
251 placeholder = '%s' |
|
252 |
|
253 if hasattr(val, 'evaluate'): |
|
254 val = SQLEvaluator(val, self, allow_joins=False) |
|
255 if model: |
|
256 self.add_related_update(model, field.column, val, placeholder) |
|
257 else: |
|
258 self.values.append((field.column, val, placeholder)) |
|
259 |
|
260 def add_related_update(self, model, column, value, placeholder): |
|
261 """ |
|
262 Adds (name, value) to an update query for an ancestor model. |
|
263 |
|
264 Updates are coalesced so that we only run one update query per ancestor. |
|
265 """ |
|
266 try: |
|
267 self.related_updates[model].append((column, value, placeholder)) |
|
268 except KeyError: |
|
269 self.related_updates[model] = [(column, value, placeholder)] |
|
270 |
|
271 def get_related_updates(self): |
|
272 """ |
|
273 Returns a list of query objects: one for each update required to an |
|
274 ancestor model. Each query will have the same filtering conditions as |
|
275 the current query but will only update a single table. |
|
276 """ |
|
277 if not self.related_updates: |
|
278 return [] |
|
279 result = [] |
|
280 for model, values in self.related_updates.iteritems(): |
|
281 query = UpdateQuery(model, self.connection) |
|
282 query.values = values |
|
283 if self.related_ids: |
|
284 query.add_filter(('pk__in', self.related_ids)) |
|
285 result.append(query) |
|
286 return result |
|
287 |
|
288 class InsertQuery(Query): |
|
289 def __init__(self, *args, **kwargs): |
|
290 super(InsertQuery, self).__init__(*args, **kwargs) |
|
291 self.columns = [] |
|
292 self.values = [] |
|
293 self.params = () |
|
294 self.return_id = False |
|
295 |
|
296 def clone(self, klass=None, **kwargs): |
|
297 extras = {'columns': self.columns[:], 'values': self.values[:], |
|
298 'params': self.params, 'return_id': self.return_id} |
|
299 extras.update(kwargs) |
|
300 return super(InsertQuery, self).clone(klass, **extras) |
|
301 |
|
302 def as_sql(self): |
|
303 # We don't need quote_name_unless_alias() here, since these are all |
|
304 # going to be column names (so we can avoid the extra overhead). |
|
305 qn = self.connection.ops.quote_name |
|
306 opts = self.model._meta |
|
307 result = ['INSERT INTO %s' % qn(opts.db_table)] |
|
308 result.append('(%s)' % ', '.join([qn(c) for c in self.columns])) |
|
309 result.append('VALUES (%s)' % ', '.join(self.values)) |
|
310 params = self.params |
|
311 if self.return_id and self.connection.features.can_return_id_from_insert: |
|
312 col = "%s.%s" % (qn(opts.db_table), qn(opts.pk.column)) |
|
313 r_fmt, r_params = self.connection.ops.return_insert_id() |
|
314 result.append(r_fmt % col) |
|
315 params = params + r_params |
|
316 return ' '.join(result), params |
|
317 |
|
318 def execute_sql(self, return_id=False): |
|
319 self.return_id = return_id |
|
320 cursor = super(InsertQuery, self).execute_sql(None) |
|
321 if not (return_id and cursor): |
|
322 return |
|
323 if self.connection.features.can_return_id_from_insert: |
|
324 return self.connection.ops.fetch_returned_insert_id(cursor) |
|
325 return self.connection.ops.last_insert_id(cursor, |
|
326 self.model._meta.db_table, self.model._meta.pk.column) |
|
327 |
|
328 def insert_values(self, insert_values, raw_values=False): |
|
329 """ |
|
330 Set up the insert query from the 'insert_values' dictionary. The |
|
331 dictionary gives the model field names and their target values. |
|
332 |
|
333 If 'raw_values' is True, the values in the 'insert_values' dictionary |
|
334 are inserted directly into the query, rather than passed as SQL |
|
335 parameters. This provides a way to insert NULL and DEFAULT keywords |
|
336 into the query, for example. |
|
337 """ |
|
338 placeholders, values = [], [] |
|
339 for field, val in insert_values: |
|
340 if hasattr(field, 'get_placeholder'): |
|
341 # Some fields (e.g. geo fields) need special munging before |
|
342 # they can be inserted. |
|
343 placeholders.append(field.get_placeholder(val)) |
|
344 else: |
|
345 placeholders.append('%s') |
|
346 |
|
347 self.columns.append(field.column) |
|
348 values.append(val) |
|
349 if raw_values: |
|
350 self.values.extend(values) |
|
351 else: |
|
352 self.params += tuple(values) |
|
353 self.values.extend(placeholders) |
|
354 |
|
355 class DateQuery(Query): |
|
356 """ |
|
357 A DateQuery is a normal query, except that it specifically selects a single |
|
358 date field. This requires some special handling when converting the results |
|
359 back to Python objects, so we put it in a separate class. |
|
360 """ |
|
361 def __getstate__(self): |
|
362 """ |
|
363 Special DateQuery-specific pickle handling. |
|
364 """ |
|
365 for elt in self.select: |
|
366 if isinstance(elt, Date): |
|
367 # Eliminate a method reference that can't be pickled. The |
|
368 # __setstate__ method restores this. |
|
369 elt.date_sql_func = None |
|
370 return super(DateQuery, self).__getstate__() |
|
371 |
|
372 def __setstate__(self, obj_dict): |
|
373 super(DateQuery, self).__setstate__(obj_dict) |
|
374 for elt in self.select: |
|
375 if isinstance(elt, Date): |
|
376 self.date_sql_func = self.connection.ops.date_trunc_sql |
|
377 |
|
378 def results_iter(self): |
|
379 """ |
|
380 Returns an iterator over the results from executing this query. |
|
381 """ |
|
382 resolve_columns = hasattr(self, 'resolve_columns') |
|
383 if resolve_columns: |
|
384 from django.db.models.fields import DateTimeField |
|
385 fields = [DateTimeField()] |
|
386 else: |
|
387 from django.db.backends.util import typecast_timestamp |
|
388 needs_string_cast = self.connection.features.needs_datetime_string_cast |
|
389 |
|
390 offset = len(self.extra_select) |
|
391 for rows in self.execute_sql(MULTI): |
|
392 for row in rows: |
|
393 date = row[offset] |
|
394 if resolve_columns: |
|
395 date = self.resolve_columns(row, fields)[offset] |
|
396 elif needs_string_cast: |
|
397 date = typecast_timestamp(str(date)) |
|
398 yield date |
|
399 |
|
400 def add_date_select(self, field, lookup_type, order='ASC'): |
|
401 """ |
|
402 Converts the query into a date extraction query. |
|
403 """ |
|
404 result = self.setup_joins([field.name], self.get_meta(), |
|
405 self.get_initial_alias(), False) |
|
406 alias = result[3][-1] |
|
407 select = Date((alias, field.column), lookup_type, |
|
408 self.connection.ops.date_trunc_sql) |
|
409 self.select = [select] |
|
410 self.select_fields = [None] |
|
411 self.select_related = False # See #7097. |
|
412 self.extra = {} |
|
413 self.distinct = True |
|
414 self.order_by = order == 'ASC' and [1] or [-1] |
|
415 |
|
416 class AggregateQuery(Query): |
|
417 """ |
|
418 An AggregateQuery takes another query as a parameter to the FROM |
|
419 clause and only selects the elements in the provided list. |
|
420 """ |
|
421 def add_subquery(self, query): |
|
422 self.subquery, self.sub_params = query.as_sql(with_col_aliases=True) |
|
423 |
|
424 def as_sql(self, quote_func=None): |
|
425 """ |
|
426 Creates the SQL for this query. Returns the SQL string and list of |
|
427 parameters. |
|
428 """ |
|
429 sql = ('SELECT %s FROM (%s) subquery' % ( |
|
430 ', '.join([ |
|
431 aggregate.as_sql() |
|
432 for aggregate in self.aggregate_select.values() |
|
433 ]), |
|
434 self.subquery) |
|
435 ) |
|
436 params = self.sub_params |
|
437 return (sql, params) |