summaryrefslogtreecommitdiffstats
path: root/tests/test_optimizer.py
blob: af216797fecc2604eee9139d498ccd1d2a9b0723 (plain)
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
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
import unittest
from functools import partial

import duckdb
from pandas.testing import assert_frame_equal

import sqlglot
from sqlglot import exp, optimizer, parse_one
from sqlglot.errors import OptimizeError, SchemaError
from sqlglot.optimizer.annotate_types import annotate_types
from sqlglot.optimizer.scope import build_scope, traverse_scope, walk_in_scope
from sqlglot.schema import MappingSchema
from tests.helpers import (
    TPCH_SCHEMA,
    load_sql_fixture_pairs,
    load_sql_fixtures,
    string_to_bool,
)


class TestOptimizer(unittest.TestCase):
    maxDiff = None

    @classmethod
    def setUpClass(cls):
        cls.conn = duckdb.connect()
        cls.conn.execute(
            """
        CREATE TABLE x (a INT, b INT);
        CREATE TABLE y (b INT, c INT);
        CREATE TABLE z (b INT, c INT);
        CREATE TABLE w (d TEXT, e TEXT);

        INSERT INTO x VALUES (1, 1);
        INSERT INTO x VALUES (2, 2);
        INSERT INTO x VALUES (2, 2);
        INSERT INTO x VALUES (3, 3);
        INSERT INTO x VALUES (null, null);

        INSERT INTO y VALUES (2, 2);
        INSERT INTO y VALUES (2, 2);
        INSERT INTO y VALUES (3, 3);
        INSERT INTO y VALUES (4, 4);
        INSERT INTO y VALUES (null, null);

        INSERT INTO y VALUES (3, 3);
        INSERT INTO y VALUES (3, 3);
        INSERT INTO y VALUES (4, 4);
        INSERT INTO y VALUES (5, 5);
        INSERT INTO y VALUES (null, null);

        INSERT INTO w VALUES ('a', 'b');
        """
        )

    def setUp(self):
        self.schema = {
            "x": {
                "a": "INT",
                "b": "INT",
            },
            "y": {
                "b": "INT",
                "c": "INT",
            },
            "z": {
                "b": "INT",
                "c": "INT",
            },
            "w": {
                "d": "TEXT",
                "e": "TEXT",
            },
        }

    def check_file(self, file, func, pretty=False, execute=False, **kwargs):
        for i, (meta, sql, expected) in enumerate(
            load_sql_fixture_pairs(f"optimizer/{file}.sql"), start=1
        ):
            title = meta.get("title") or f"{i}, {sql}"
            dialect = meta.get("dialect")
            leave_tables_isolated = meta.get("leave_tables_isolated")

            func_kwargs = {**kwargs}
            if leave_tables_isolated is not None:
                func_kwargs["leave_tables_isolated"] = string_to_bool(leave_tables_isolated)

            with self.subTest(title):
                optimized = func(parse_one(sql, read=dialect), **func_kwargs)
                self.assertEqual(
                    expected,
                    optimized.sql(pretty=pretty, dialect=dialect),
                )

            should_execute = meta.get("execute")
            if should_execute is None:
                should_execute = execute

            if string_to_bool(should_execute):
                with self.subTest(f"(execute) {title}"):
                    df1 = self.conn.execute(
                        sqlglot.transpile(sql, read=dialect, write="duckdb")[0]
                    ).df()
                    df2 = self.conn.execute(optimized.sql(pretty=pretty, dialect="duckdb")).df()
                    assert_frame_equal(df1, df2)

    def test_optimize(self):
        schema = {
            "x": {"a": "INT", "b": "INT"},
            "y": {"b": "INT", "c": "INT"},
            "z": {"a": "INT", "c": "INT"},
        }

        self.check_file("optimizer", optimizer.optimize, pretty=True, execute=True, schema=schema)

    def test_isolate_table_selects(self):
        self.check_file(
            "isolate_table_selects",
            optimizer.isolate_table_selects.isolate_table_selects,
        )

    def test_qualify_tables(self):
        self.check_file(
            "qualify_tables",
            optimizer.qualify_tables.qualify_tables,
            db="db",
            catalog="c",
        )

    def test_normalize(self):
        self.assertEqual(
            optimizer.normalize.normalize(
                parse_one("x AND (y OR z)"),
                dnf=True,
            ).sql(),
            "(x AND y) OR (x AND z)",
        )

        self.check_file(
            "normalize",
            optimizer.normalize.normalize,
        )

    def test_qualify_columns(self):
        def qualify_columns(expression, **kwargs):
            expression = optimizer.qualify_tables.qualify_tables(expression)
            expression = optimizer.qualify_columns.qualify_columns(expression, **kwargs)
            return expression

        self.check_file("qualify_columns", qualify_columns, execute=True, schema=self.schema)

    def test_qualify_columns__with_invisible(self):
        def qualify_columns(expression, **kwargs):
            expression = optimizer.qualify_tables.qualify_tables(expression)
            expression = optimizer.qualify_columns.qualify_columns(expression, **kwargs)
            return expression

        schema = MappingSchema(self.schema, {"x": {"a"}, "y": {"b"}, "z": {"b"}})
        self.check_file("qualify_columns__with_invisible", qualify_columns, schema=schema)

    def test_qualify_columns__invalid(self):
        for sql in load_sql_fixtures("optimizer/qualify_columns__invalid.sql"):
            with self.subTest(sql):
                with self.assertRaises((OptimizeError, SchemaError)):
                    optimizer.qualify_columns.qualify_columns(parse_one(sql), schema=self.schema)

    def test_lower_identities(self):
        self.check_file("lower_identities", optimizer.lower_identities.lower_identities)

    def test_pushdown_projection(self):
        def pushdown_projections(expression, **kwargs):
            expression = optimizer.qualify_tables.qualify_tables(expression)
            expression = optimizer.qualify_columns.qualify_columns(expression, **kwargs)
            expression = optimizer.pushdown_projections.pushdown_projections(expression)
            return expression

        self.check_file("pushdown_projections", pushdown_projections, schema=self.schema)

    def test_simplify(self):
        self.check_file("simplify", optimizer.simplify.simplify)

    def test_unnest_subqueries(self):
        self.check_file(
            "unnest_subqueries",
            optimizer.unnest_subqueries.unnest_subqueries,
            pretty=True,
        )

    def test_pushdown_predicates(self):
        self.check_file("pushdown_predicates", optimizer.pushdown_predicates.pushdown_predicates)

    def test_expand_multi_table_selects(self):
        self.check_file(
            "expand_multi_table_selects",
            optimizer.expand_multi_table_selects.expand_multi_table_selects,
        )

    def test_optimize_joins(self):
        self.check_file(
            "optimize_joins",
            optimizer.optimize_joins.optimize_joins,
        )

    def test_eliminate_joins(self):
        self.check_file(
            "eliminate_joins",
            optimizer.eliminate_joins.eliminate_joins,
            pretty=True,
        )

    def test_eliminate_ctes(self):
        self.check_file(
            "eliminate_ctes",
            optimizer.eliminate_ctes.eliminate_ctes,
            pretty=True,
        )

    def test_merge_subqueries(self):
        optimize = partial(
            optimizer.optimize,
            rules=[
                optimizer.qualify_tables.qualify_tables,
                optimizer.qualify_columns.qualify_columns,
                optimizer.merge_subqueries.merge_subqueries,
            ],
        )

        self.check_file("merge_subqueries", optimize, execute=True, schema=self.schema)

    def test_eliminate_subqueries(self):
        self.check_file("eliminate_subqueries", optimizer.eliminate_subqueries.eliminate_subqueries)

    def test_canonicalize(self):
        optimize = partial(
            optimizer.optimize,
            rules=[
                optimizer.qualify_tables.qualify_tables,
                optimizer.qualify_columns.qualify_columns,
                annotate_types,
                optimizer.canonicalize.canonicalize,
            ],
        )
        self.check_file("canonicalize", optimize, schema=self.schema)

    def test_tpch(self):
        self.check_file("tpc-h/tpc-h", optimizer.optimize, schema=TPCH_SCHEMA, pretty=True)

    def test_file_schema(self):
        expression = parse_one(
            """
            SELECT *
            FROM READ_CSV('tests/fixtures/optimizer/tpc-h/nation.csv.gz', 'delimiter', '|')
            """
        )
        self.assertEqual(
            """
SELECT
  "_q_0"."n_nationkey" AS "n_nationkey",
  "_q_0"."n_name" AS "n_name",
  "_q_0"."n_regionkey" AS "n_regionkey",
  "_q_0"."n_comment" AS "n_comment"
FROM READ_CSV('tests/fixtures/optimizer/tpc-h/nation.csv.gz', 'delimiter', '|') AS "_q_0"
""".strip(),
            optimizer.optimize(expression).sql(pretty=True),
        )

    def test_scope(self):
        sql = """
        WITH q AS (
          SELECT x.b FROM x
        ), r AS (
          SELECT y.b FROM y
        ), z as (
          SELECT cola, colb FROM (VALUES(1, 'test')) AS tab(cola, colb)
        )
        SELECT
          r.b,
          s.b
        FROM r
        JOIN (
          SELECT y.c AS b FROM y
        ) s
        ON s.b = r.b
        WHERE s.b > (SELECT MAX(x.a) FROM x WHERE x.b = s.b)
        """
        expression = parse_one(sql)
        for scopes in traverse_scope(expression), list(build_scope(expression).traverse()):
            self.assertEqual(len(scopes), 7)
            self.assertEqual(scopes[0].expression.sql(), "SELECT x.b FROM x")
            self.assertEqual(scopes[1].expression.sql(), "SELECT y.b FROM y")
            self.assertEqual(scopes[2].expression.sql(), "(VALUES (1, 'test')) AS tab(cola, colb)")
            self.assertEqual(
                scopes[3].expression.sql(),
                "SELECT cola, colb FROM (VALUES (1, 'test')) AS tab(cola, colb)",
            )
            self.assertEqual(scopes[4].expression.sql(), "SELECT y.c AS b FROM y")
            self.assertEqual(scopes[5].expression.sql(), "SELECT MAX(x.a) FROM x WHERE x.b = s.b")
            self.assertEqual(scopes[6].expression.sql(), parse_one(sql).sql())

            self.assertEqual(set(scopes[6].sources), {"q", "z", "r", "s"})
            self.assertEqual(len(scopes[6].columns), 6)
            self.assertEqual({c.table for c in scopes[6].columns}, {"r", "s"})
            self.assertEqual(scopes[6].source_columns("q"), [])
            self.assertEqual(len(scopes[6].source_columns("r")), 2)
            self.assertEqual({c.table for c in scopes[6].source_columns("r")}, {"r"})

            self.assertEqual({c.sql() for c in scopes[-1].find_all(exp.Column)}, {"r.b", "s.b"})
            self.assertEqual(scopes[-1].find(exp.Column).sql(), "r.b")
            self.assertEqual({c.sql() for c in scopes[0].find_all(exp.Column)}, {"x.b"})

        # Check that we can walk in scope from an arbitrary node
        self.assertEqual(
            {
                node.sql()
                for node, *_ in walk_in_scope(expression.find(exp.Where))
                if isinstance(node, exp.Column)
            },
            {"s.b"},
        )

    def test_literal_type_annotation(self):
        tests = {
            "SELECT 5": exp.DataType.Type.INT,
            "SELECT 5.3": exp.DataType.Type.DOUBLE,
            "SELECT 'bla'": exp.DataType.Type.VARCHAR,
            "5": exp.DataType.Type.INT,
            "5.3": exp.DataType.Type.DOUBLE,
            "'bla'": exp.DataType.Type.VARCHAR,
        }

        for sql, target_type in tests.items():
            expression = annotate_types(parse_one(sql))
            self.assertEqual(expression.find(exp.Literal).type.this, target_type)

    def test_boolean_type_annotation(self):
        tests = {
            "SELECT TRUE": exp.DataType.Type.BOOLEAN,
            "FALSE": exp.DataType.Type.BOOLEAN,
        }

        for sql, target_type in tests.items():
            expression = annotate_types(parse_one(sql))
            self.assertEqual(expression.find(exp.Boolean).type.this, target_type)

    def test_cast_type_annotation(self):
        expression = annotate_types(parse_one("CAST('2020-01-01' AS TIMESTAMPTZ(9))"))
        self.assertEqual(expression.type.this, exp.DataType.Type.TIMESTAMPTZ)
        self.assertEqual(expression.this.type.this, exp.DataType.Type.VARCHAR)
        self.assertEqual(expression.args["to"].type.this, exp.DataType.Type.TIMESTAMPTZ)
        self.assertEqual(expression.args["to"].expressions[0].type.this, exp.DataType.Type.INT)

        expression = annotate_types(parse_one("ARRAY(1)::ARRAY<INT>"))
        self.assertEqual(expression.type, parse_one("ARRAY<INT>", into=exp.DataType))

    def test_cache_annotation(self):
        expression = annotate_types(
            parse_one("CACHE LAZY TABLE x OPTIONS('storageLevel' = 'value') AS SELECT 1")
        )
        self.assertEqual(expression.expression.expressions[0].type.this, exp.DataType.Type.INT)

    def test_binary_annotation(self):
        expression = annotate_types(parse_one("SELECT 0.0 + (2 + 3)")).expressions[0]

        self.assertEqual(expression.type.this, exp.DataType.Type.DOUBLE)
        self.assertEqual(expression.left.type.this, exp.DataType.Type.DOUBLE)
        self.assertEqual(expression.right.type.this, exp.DataType.Type.INT)
        self.assertEqual(expression.right.this.type.this, exp.DataType.Type.INT)
        self.assertEqual(expression.right.this.left.type.this, exp.DataType.Type.INT)
        self.assertEqual(expression.right.this.right.type.this, exp.DataType.Type.INT)

    def test_derived_tables_column_annotation(self):
        schema = {"x": {"cola": "INT"}, "y": {"cola": "FLOAT"}}
        sql = """
            SELECT a.cola AS cola
            FROM (
                SELECT x.cola + y.cola AS cola
                FROM (
                    SELECT x.cola AS cola
                    FROM x AS x
                ) AS x
                JOIN (
                    SELECT y.cola AS cola
                    FROM y AS y
                ) AS y
            ) AS a
        """

        expression = annotate_types(parse_one(sql), schema=schema)
        self.assertEqual(
            expression.expressions[0].type.this, exp.DataType.Type.FLOAT
        )  # a.cola AS cola

        addition_alias = expression.args["from"].expressions[0].this.expressions[0]
        self.assertEqual(
            addition_alias.type.this, exp.DataType.Type.FLOAT
        )  # x.cola + y.cola AS cola

        addition = addition_alias.this
        self.assertEqual(addition.type.this, exp.DataType.Type.FLOAT)
        self.assertEqual(addition.this.type.this, exp.DataType.Type.INT)
        self.assertEqual(addition.expression.type.this, exp.DataType.Type.FLOAT)

    def test_cte_column_annotation(self):
        schema = {"x": {"cola": "CHAR"}, "y": {"colb": "TEXT", "colc": "BOOLEAN"}}
        sql = """
            WITH tbl AS (
                SELECT x.cola + 'bla' AS cola, y.colb AS colb, y.colc AS colc
                FROM (
                    SELECT x.cola AS cola
                    FROM x AS x
                ) AS x
                JOIN (
                    SELECT y.colb AS colb, y.colc AS colc
                    FROM y AS y
                ) AS y
            )
            SELECT tbl.cola + tbl.colb + 'foo' AS col
            FROM tbl AS tbl
            WHERE tbl.colc = True
        """

        expression = annotate_types(parse_one(sql), schema=schema)
        self.assertEqual(
            expression.expressions[0].type.this, exp.DataType.Type.TEXT
        )  # tbl.cola + tbl.colb + 'foo' AS col

        outer_addition = expression.expressions[0].this  # (tbl.cola + tbl.colb) + 'foo'
        self.assertEqual(outer_addition.type.this, exp.DataType.Type.TEXT)
        self.assertEqual(outer_addition.left.type.this, exp.DataType.Type.TEXT)
        self.assertEqual(outer_addition.right.type.this, exp.DataType.Type.VARCHAR)

        inner_addition = expression.expressions[0].this.left  # tbl.cola + tbl.colb
        self.assertEqual(inner_addition.left.type.this, exp.DataType.Type.VARCHAR)
        self.assertEqual(inner_addition.right.type.this, exp.DataType.Type.TEXT)

        # WHERE tbl.colc = True
        self.assertEqual(expression.args["where"].this.type.this, exp.DataType.Type.BOOLEAN)

        cte_select = expression.args["with"].expressions[0].this
        self.assertEqual(
            cte_select.expressions[0].type.this, exp.DataType.Type.VARCHAR
        )  # x.cola + 'bla' AS cola
        self.assertEqual(
            cte_select.expressions[1].type.this, exp.DataType.Type.TEXT
        )  # y.colb AS colb
        self.assertEqual(
            cte_select.expressions[2].type.this, exp.DataType.Type.BOOLEAN
        )  # y.colc AS colc

        cte_select_addition = cte_select.expressions[0].this  # x.cola + 'bla'
        self.assertEqual(cte_select_addition.type.this, exp.DataType.Type.VARCHAR)
        self.assertEqual(cte_select_addition.left.type.this, exp.DataType.Type.CHAR)
        self.assertEqual(cte_select_addition.right.type.this, exp.DataType.Type.VARCHAR)

        # Check that x.cola AS cola and y.colb AS colb have types CHAR and TEXT, respectively
        for d, t in zip(
            cte_select.find_all(exp.Subquery), [exp.DataType.Type.CHAR, exp.DataType.Type.TEXT]
        ):
            self.assertEqual(d.this.expressions[0].this.type.this, t)

    def test_function_annotation(self):
        schema = {"x": {"cola": "VARCHAR", "colb": "CHAR"}}
        sql = "SELECT x.cola || TRIM(x.colb) AS col FROM x AS x"

        concat_expr_alias = annotate_types(parse_one(sql), schema=schema).expressions[0]
        self.assertEqual(concat_expr_alias.type.this, exp.DataType.Type.VARCHAR)

        concat_expr = concat_expr_alias.this
        self.assertEqual(concat_expr.type.this, exp.DataType.Type.VARCHAR)
        self.assertEqual(concat_expr.left.type.this, exp.DataType.Type.VARCHAR)  # x.cola
        self.assertEqual(concat_expr.right.type.this, exp.DataType.Type.VARCHAR)  # TRIM(x.colb)
        self.assertEqual(concat_expr.right.this.type.this, exp.DataType.Type.CHAR)  # x.colb

        sql = "SELECT CASE WHEN 1=1 THEN x.cola ELSE x.colb END AS col FROM x AS x"

        case_expr_alias = annotate_types(parse_one(sql), schema=schema).expressions[0]
        self.assertEqual(case_expr_alias.type.this, exp.DataType.Type.VARCHAR)

        case_expr = case_expr_alias.this
        self.assertEqual(case_expr.type.this, exp.DataType.Type.VARCHAR)
        self.assertEqual(case_expr.args["default"].type.this, exp.DataType.Type.CHAR)

        case_ifs_expr = case_expr.args["ifs"][0]
        self.assertEqual(case_ifs_expr.type.this, exp.DataType.Type.VARCHAR)
        self.assertEqual(case_ifs_expr.args["true"].type.this, exp.DataType.Type.VARCHAR)

    def test_unknown_annotation(self):
        schema = {"x": {"cola": "VARCHAR"}}
        sql = "SELECT x.cola || SOME_ANONYMOUS_FUNC(x.cola) AS col FROM x AS x"

        concat_expr_alias = annotate_types(parse_one(sql), schema=schema).expressions[0]
        self.assertEqual(concat_expr_alias.type.this, exp.DataType.Type.UNKNOWN)

        concat_expr = concat_expr_alias.this
        self.assertEqual(concat_expr.type.this, exp.DataType.Type.UNKNOWN)
        self.assertEqual(concat_expr.left.type.this, exp.DataType.Type.VARCHAR)  # x.cola
        self.assertEqual(
            concat_expr.right.type.this, exp.DataType.Type.UNKNOWN
        )  # SOME_ANONYMOUS_FUNC(x.cola)
        self.assertEqual(
            concat_expr.right.expressions[0].type.this, exp.DataType.Type.VARCHAR
        )  # x.cola (arg)

    def test_null_annotation(self):
        expression = annotate_types(parse_one("SELECT NULL + 2 AS col")).expressions[0].this
        self.assertEqual(expression.left.type.this, exp.DataType.Type.NULL)
        self.assertEqual(expression.right.type.this, exp.DataType.Type.INT)

        # NULL <op> UNKNOWN should yield NULL
        sql = "SELECT NULL || SOME_ANONYMOUS_FUNC() AS result"

        concat_expr_alias = annotate_types(parse_one(sql)).expressions[0]
        self.assertEqual(concat_expr_alias.type.this, exp.DataType.Type.NULL)

        concat_expr = concat_expr_alias.this
        self.assertEqual(concat_expr.type.this, exp.DataType.Type.NULL)
        self.assertEqual(concat_expr.left.type.this, exp.DataType.Type.NULL)
        self.assertEqual(concat_expr.right.type.this, exp.DataType.Type.UNKNOWN)

    def test_nullable_annotation(self):
        nullable = exp.DataType.build("NULLABLE", expressions=exp.DataType.build("BOOLEAN"))
        expression = annotate_types(parse_one("NULL AND FALSE"))

        self.assertEqual(expression.type, nullable)
        self.assertEqual(expression.left.type.this, exp.DataType.Type.NULL)
        self.assertEqual(expression.right.type.this, exp.DataType.Type.BOOLEAN)

    def test_predicate_annotation(self):
        expression = annotate_types(parse_one("x BETWEEN a AND b"))
        self.assertEqual(expression.type.this, exp.DataType.Type.BOOLEAN)

        expression = annotate_types(parse_one("x IN (a, b, c, d)"))
        self.assertEqual(expression.type.this, exp.DataType.Type.BOOLEAN)

    def test_aggfunc_annotation(self):
        schema = {"x": {"cola": "SMALLINT", "colb": "FLOAT", "colc": "TEXT", "cold": "DATE"}}

        tests = {
            ("AVG", "cola"): exp.DataType.Type.DOUBLE,
            ("SUM", "cola"): exp.DataType.Type.BIGINT,
            ("SUM", "colb"): exp.DataType.Type.DOUBLE,
            ("MIN", "cola"): exp.DataType.Type.SMALLINT,
            ("MIN", "colb"): exp.DataType.Type.FLOAT,
            ("MAX", "colc"): exp.DataType.Type.TEXT,
            ("MAX", "cold"): exp.DataType.Type.DATE,
            ("COUNT", "colb"): exp.DataType.Type.BIGINT,
            ("STDDEV", "cola"): exp.DataType.Type.DOUBLE,
        }

        for (func, col), target_type in tests.items():
            expression = annotate_types(
                parse_one(f"SELECT {func}(x.{col}) AS _col_0 FROM x AS x"), schema=schema
            )
            self.assertEqual(expression.expressions[0].type.this, target_type)

    def test_recursive_cte(self):
        query = parse_one(
            """
            with recursive t(n) AS
            (
              select 1
              union all
              select n + 1
              FROM t
              where n < 3
            ), y AS (
              select n
              FROM t
              union all
              select n + 1
              FROM y
              where n < 2
            )
            select * from y
            """
        )

        scope_t, scope_y = build_scope(query).cte_scopes
        self.assertEqual(set(scope_t.cte_sources), {"t"})
        self.assertEqual(set(scope_y.cte_sources), {"t", "y"})

    def test_schema_with_spaces(self):
        schema = {
            "a": {
                "b c": "text",
                '"d e"': "text",
            }
        }

        self.assertEqual(
            optimizer.optimize(parse_one("SELECT * FROM a"), schema=schema),
            parse_one('SELECT "a"."b c" AS "b c", "a"."d e" AS "d e" FROM "a" AS "a"'),
        )