1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
|
from __future__ import unicode_literals
from datetime import date
from django.db.models.query_utils import InvalidQuery
from django.test import TestCase, skipUnlessDBFeature
from .models import Author, Book, Coffee, FriendlyAuthor, Reviewer
class RawQueryTests(TestCase):
fixtures = ['raw_query_books.json']
def assertSuccessfulRawQuery(self, model, query, expected_results,
expected_annotations=(), params=[], translations=None):
"""
Execute the passed query against the passed model and check the output
"""
results = list(model.objects.raw(query, params=params, translations=translations))
self.assertProcessed(model, results, expected_results, expected_annotations)
self.assertAnnotations(results, expected_annotations)
def assertProcessed(self, model, results, orig, expected_annotations=()):
"""
Compare the results of a raw query against expected results
"""
self.assertEqual(len(results), len(orig))
for index, item in enumerate(results):
orig_item = orig[index]
for annotation in expected_annotations:
setattr(orig_item, *annotation)
for field in model._meta.fields:
# Check that all values on the model are equal
self.assertEqual(
getattr(item, field.attname),
getattr(orig_item, field.attname)
)
# This includes checking that they are the same type
self.assertEqual(
type(getattr(item, field.attname)),
type(getattr(orig_item, field.attname))
)
def assertNoAnnotations(self, results):
"""
Check that the results of a raw query contain no annotations
"""
self.assertAnnotations(results, ())
def assertAnnotations(self, results, expected_annotations):
"""
Check that the passed raw query results contain the expected
annotations
"""
if expected_annotations:
for index, result in enumerate(results):
annotation, value = expected_annotations[index]
self.assertTrue(hasattr(result, annotation))
self.assertEqual(getattr(result, annotation), value)
def test_simple_raw_query(self):
"""
Basic test of raw query with a simple database query
"""
query = "SELECT * FROM raw_query_author"
authors = Author.objects.all()
self.assertSuccessfulRawQuery(Author, query, authors)
def test_raw_query_lazy(self):
"""
Raw queries are lazy: they aren't actually executed until they're
iterated over.
"""
q = Author.objects.raw('SELECT * FROM raw_query_author')
self.assertIsNone(q.query.cursor)
list(q)
self.assertIsNotNone(q.query.cursor)
def test_FK_raw_query(self):
"""
Test of a simple raw query against a model containing a foreign key
"""
query = "SELECT * FROM raw_query_book"
books = Book.objects.all()
self.assertSuccessfulRawQuery(Book, query, books)
def test_db_column_handler(self):
"""
Test of a simple raw query against a model containing a field with
db_column defined.
"""
query = "SELECT * FROM raw_query_coffee"
coffees = Coffee.objects.all()
self.assertSuccessfulRawQuery(Coffee, query, coffees)
def test_order_handler(self):
"""
Test of raw raw query's tolerance for columns being returned in any
order
"""
selects = (
('dob, last_name, first_name, id'),
('last_name, dob, first_name, id'),
('first_name, last_name, dob, id'),
)
for select in selects:
query = "SELECT %s FROM raw_query_author" % select
authors = Author.objects.all()
self.assertSuccessfulRawQuery(Author, query, authors)
def test_translations(self):
"""
Test of raw query's optional ability to translate unexpected result
column names to specific model fields
"""
query = "SELECT first_name AS first, last_name AS last, dob, id FROM raw_query_author"
translations = {'first': 'first_name', 'last': 'last_name'}
authors = Author.objects.all()
self.assertSuccessfulRawQuery(Author, query, authors, translations=translations)
def test_params(self):
"""
Test passing optional query parameters
"""
query = "SELECT * FROM raw_query_author WHERE first_name = %s"
author = Author.objects.all()[2]
params = [author.first_name]
qset = Author.objects.raw(query, params=params)
results = list(qset)
self.assertProcessed(Author, results, [author])
self.assertNoAnnotations(results)
self.assertEqual(len(results), 1)
self.assertIsInstance(repr(qset), str)
@skipUnlessDBFeature('supports_paramstyle_pyformat')
def test_pyformat_params(self):
"""
Test passing optional query parameters
"""
query = "SELECT * FROM raw_query_author WHERE first_name = %(first)s"
author = Author.objects.all()[2]
params = {'first': author.first_name}
qset = Author.objects.raw(query, params=params)
results = list(qset)
self.assertProcessed(Author, results, [author])
self.assertNoAnnotations(results)
self.assertEqual(len(results), 1)
self.assertIsInstance(repr(qset), str)
def test_query_representation(self):
"""
Test representation of raw query with parameters
"""
query = "SELECT * FROM raw_query_author WHERE last_name = %(last)s"
qset = Author.objects.raw(query, {'last': 'foo'})
self.assertEqual(repr(qset), "<RawQuerySet: SELECT * FROM raw_query_author WHERE last_name = foo>")
self.assertEqual(repr(qset.query), "<RawQuery: SELECT * FROM raw_query_author WHERE last_name = foo>")
query = "SELECT * FROM raw_query_author WHERE last_name = %s"
qset = Author.objects.raw(query, {'foo'})
self.assertEqual(repr(qset), "<RawQuerySet: SELECT * FROM raw_query_author WHERE last_name = foo>")
self.assertEqual(repr(qset.query), "<RawQuery: SELECT * FROM raw_query_author WHERE last_name = foo>")
def test_many_to_many(self):
"""
Test of a simple raw query against a model containing a m2m field
"""
query = "SELECT * FROM raw_query_reviewer"
reviewers = Reviewer.objects.all()
self.assertSuccessfulRawQuery(Reviewer, query, reviewers)
def test_extra_conversions(self):
"""
Test to insure that extra translations are ignored.
"""
query = "SELECT * FROM raw_query_author"
translations = {'something': 'else'}
authors = Author.objects.all()
self.assertSuccessfulRawQuery(Author, query, authors, translations=translations)
def test_missing_fields(self):
query = "SELECT id, first_name, dob FROM raw_query_author"
for author in Author.objects.raw(query):
self.assertNotEqual(author.first_name, None)
# last_name isn't given, but it will be retrieved on demand
self.assertNotEqual(author.last_name, None)
def test_missing_fields_without_PK(self):
query = "SELECT first_name, dob FROM raw_query_author"
try:
list(Author.objects.raw(query))
self.fail('Query without primary key should fail')
except InvalidQuery:
pass
def test_annotations(self):
query = "SELECT a.*, count(b.id) as book_count FROM raw_query_author a LEFT JOIN raw_query_book b ON a.id = b.author_id GROUP BY a.id, a.first_name, a.last_name, a.dob ORDER BY a.id"
expected_annotations = (
('book_count', 3),
('book_count', 0),
('book_count', 1),
('book_count', 0),
)
authors = Author.objects.all()
self.assertSuccessfulRawQuery(Author, query, authors, expected_annotations)
def test_white_space_query(self):
query = " SELECT * FROM raw_query_author"
authors = Author.objects.all()
self.assertSuccessfulRawQuery(Author, query, authors)
def test_multiple_iterations(self):
query = "SELECT * FROM raw_query_author"
normal_authors = Author.objects.all()
raw_authors = Author.objects.raw(query)
# First Iteration
first_iterations = 0
for index, raw_author in enumerate(raw_authors):
self.assertEqual(normal_authors[index], raw_author)
first_iterations += 1
# Second Iteration
second_iterations = 0
for index, raw_author in enumerate(raw_authors):
self.assertEqual(normal_authors[index], raw_author)
second_iterations += 1
self.assertEqual(first_iterations, second_iterations)
def test_get_item(self):
# Indexing on RawQuerySets
query = "SELECT * FROM raw_query_author ORDER BY id ASC"
third_author = Author.objects.raw(query)[2]
self.assertEqual(third_author.first_name, 'Bob')
first_two = Author.objects.raw(query)[0:2]
self.assertEqual(len(first_two), 2)
self.assertRaises(TypeError, lambda: Author.objects.raw(query)['test'])
def test_inheritance(self):
# date is the end of the Cuban Missile Crisis, I have no idea when
# Wesley was born
f = FriendlyAuthor.objects.create(first_name="Wesley", last_name="Chun",
dob=date(1962, 10, 28))
query = "SELECT * FROM raw_query_friendlyauthor"
self.assertEqual(
[o.pk for o in FriendlyAuthor.objects.raw(query)], [f.pk]
)
def test_query_count(self):
self.assertNumQueries(1, list, Author.objects.raw("SELECT * FROM raw_query_author"))
def test_subquery_in_raw_sql(self):
try:
list(Book.objects.raw('SELECT id FROM (SELECT * FROM raw_query_book WHERE paperback IS NOT NULL) sq'))
except InvalidQuery:
self.fail("Using a subquery in a RawQuerySet raised InvalidQuery")
|