from sqlglot import exp, parse_one from tests.dialects.test_dialect import Validator class TestClickhouse(Validator): dialect = "clickhouse" def test_clickhouse(self): self.validate_identity("SELECT * FROM x LIMIT 1 UNION ALL SELECT * FROM y") string_types = [ "BLOB", "LONGBLOB", "LONGTEXT", "MEDIUMBLOB", "MEDIUMTEXT", "TINYBLOB", "TINYTEXT", "VARCHAR(255)", ] for string_type in string_types: self.validate_identity(f"CAST(x AS {string_type})", "CAST(x AS String)") expr = parse_one("count(x)") self.assertEqual(expr.sql(dialect="clickhouse"), "COUNT(x)") self.assertIsNone(expr._meta) self.validate_identity("countIf(x, y)") self.validate_identity("x = y") self.validate_identity("x <> y") self.validate_identity("SELECT * FROM (SELECT a FROM b SAMPLE 0.01)") self.validate_identity("SELECT * FROM (SELECT a FROM b SAMPLE 1 / 10 OFFSET 1 / 2)") self.validate_identity("SELECT sum(foo * bar) FROM bla SAMPLE 10000000") self.validate_identity("CAST(x AS Nested(ID UInt32, Serial UInt32, EventTime DATETIME))") self.validate_identity("CAST(x AS Enum('hello' = 1, 'world' = 2))") self.validate_identity("CAST(x AS Enum('hello', 'world'))") self.validate_identity("CAST(x AS Enum('hello' = 1, 'world'))") self.validate_identity("CAST(x AS Enum8('hello' = -123, 'world'))") self.validate_identity("CAST(x AS FixedString(1))") self.validate_identity("CAST(x AS LowCardinality(FixedString))") self.validate_identity("SELECT isNaN(1.0)") self.validate_identity("SELECT startsWith('Spider-Man', 'Spi')") self.validate_identity("SELECT xor(TRUE, FALSE)") self.validate_identity("CAST(['hello'], 'Array(Enum8(''hello'' = 1))')") self.validate_identity("SELECT x, COUNT() FROM y GROUP BY x WITH TOTALS") self.validate_identity("SELECT INTERVAL t.days DAY") self.validate_identity("SELECT match('abc', '([a-z]+)')") self.validate_identity("dictGet(x, 'y')") self.validate_identity("WITH final AS (SELECT 1) SELECT * FROM final") self.validate_identity("SELECT * FROM x FINAL") self.validate_identity("SELECT * FROM x AS y FINAL") self.validate_identity("'a' IN mapKeys(map('a', 1, 'b', 2))") self.validate_identity("CAST((1, 2) AS Tuple(a Int8, b Int16))") self.validate_identity("SELECT * FROM foo LEFT ANY JOIN bla") self.validate_identity("SELECT * FROM foo LEFT ASOF JOIN bla") self.validate_identity("SELECT * FROM foo ASOF JOIN bla") self.validate_identity("SELECT * FROM foo ANY JOIN bla") self.validate_identity("SELECT * FROM foo GLOBAL ANY JOIN bla") self.validate_identity("SELECT * FROM foo GLOBAL LEFT ANY JOIN bla") self.validate_identity("SELECT quantile(0.5)(a)") self.validate_identity("SELECT quantiles(0.5)(a) AS x FROM t") self.validate_identity("SELECT quantilesIf(0.5)(a, a > 1) AS x FROM t") self.validate_identity("SELECT quantileState(0.5)(a) AS x FROM t") self.validate_identity("SELECT deltaSumMerge(a) AS x FROM t") self.validate_identity("SELECT quantiles(0.1, 0.2, 0.3)(a)") self.validate_identity("SELECT quantileTiming(0.5)(RANGE(100))") self.validate_identity("SELECT histogram(5)(a)") self.validate_identity("SELECT groupUniqArray(2)(a)") self.validate_identity("SELECT exponentialTimeDecayedAvg(60)(a, b)") self.validate_identity("SELECT * FROM foo WHERE x GLOBAL IN (SELECT * FROM bar)") self.validate_identity("position(haystack, needle)") self.validate_identity("position(haystack, needle, position)") self.validate_identity("CAST(x AS DATETIME)") self.validate_identity("CAST(x as MEDIUMINT)", "CAST(x AS Int32)") self.validate_identity("SELECT arrayJoin([1, 2, 3] AS src) AS dst, 'Hello', src") self.validate_identity("""SELECT JSONExtractString('{"x": {"y": 1}}', 'x', 'y')""") self.validate_identity("SELECT * FROM table LIMIT 1 BY a, b") self.validate_identity("SELECT * FROM table LIMIT 2 OFFSET 1 BY a, b") self.validate_identity( "SELECT $1$foo$1$", "SELECT 'foo'", ) self.validate_identity( "SELECT * FROM table LIMIT 1, 2 BY a, b", "SELECT * FROM table LIMIT 2 OFFSET 1 BY a, b", ) self.validate_identity( "SELECT * FROM table LIMIT 1 BY CONCAT(datalayerVariantNo, datalayerProductId, warehouse)" ) self.validate_identity( """SELECT JSONExtractString('{"a": "hello", "b": [-100, 200.0, 300]}', 'a')""" ) self.validate_identity( "ATTACH DATABASE DEFAULT ENGINE = ORDINARY", check_command_warning=True ) self.validate_identity( "SELECT n, source FROM (SELECT toFloat32(number % 10) AS n, 'original' AS source FROM numbers(10) WHERE number % 3 = 1) ORDER BY n WITH FILL" ) self.validate_identity( "SELECT n, source FROM (SELECT toFloat32(number % 10) AS n, 'original' AS source FROM numbers(10) WHERE number % 3 = 1) ORDER BY n WITH FILL FROM 0 TO 5.51 STEP 0.5" ) self.validate_identity( "SELECT toDate((number * 10) * 86400) AS d1, toDate(number * 86400) AS d2, 'original' AS source FROM numbers(10) WHERE (number % 3) = 1 ORDER BY d2 WITH FILL, d1 WITH FILL STEP 5" ) self.validate_identity( "SELECT n, source, inter FROM (SELECT toFloat32(number % 10) AS n, 'original' AS source, number AS inter FROM numbers(10) WHERE number % 3 = 1) ORDER BY n WITH FILL FROM 0 TO 5.51 STEP 0.5 INTERPOLATE (inter AS inter + 1)" ) self.validate_identity( "SELECT SUM(1) AS impressions, arrayJoin(cities) AS city, arrayJoin(browsers) AS browser FROM (SELECT ['Istanbul', 'Berlin', 'Bobruisk'] AS cities, ['Firefox', 'Chrome', 'Chrome'] AS browsers) GROUP BY 2, 3" ) self.validate_identity( "SELECT sum(1) AS impressions, (arrayJoin(arrayZip(cities, browsers)) AS t).1 AS city, t.2 AS browser FROM (SELECT ['Istanbul', 'Berlin', 'Bobruisk'] AS cities, ['Firefox', 'Chrome', 'Chrome'] AS browsers) GROUP BY 2, 3" ) self.validate_identity( "SELECT SUM(1) AS impressions FROM (SELECT ['Istanbul', 'Berlin', 'Bobruisk'] AS cities) WHERE arrayJoin(cities) IN ['Istanbul', 'Berlin']", "SELECT SUM(1) AS impressions FROM (SELECT ['Istanbul', 'Berlin', 'Bobruisk'] AS cities) WHERE arrayJoin(cities) IN ('Istanbul', 'Berlin')", ) self.validate_identity( 'SELECT CAST(tuple(1 AS "a", 2 AS "b", 3.0 AS "c").2 AS Nullable(String))' ) self.validate_identity( "CREATE TABLE test (id UInt8) ENGINE=AggregatingMergeTree() ORDER BY tuple()" ) self.validate_identity( "CREATE TABLE test ON CLUSTER default (id UInt8) ENGINE=AggregatingMergeTree() ORDER BY tuple()" ) self.validate_identity( "CREATE MATERIALIZED VIEW test_view ON CLUSTER cl1 (id UInt8) ENGINE=AggregatingMergeTree() ORDER BY tuple() AS SELECT * FROM test_data" ) self.validate_identity( "CREATE MATERIALIZED VIEW test_view ON CLUSTER cl1 (id UInt8) TO table1 AS SELECT * FROM test_data" ) self.validate_identity( "CREATE MATERIALIZED VIEW test_view (id UInt8) TO db.table1 AS SELECT * FROM test_data" ) self.validate_all( "SELECT arrayJoin([1,2,3])", write={ "clickhouse": "SELECT arrayJoin([1, 2, 3])", "postgres": "SELECT UNNEST(ARRAY[1, 2, 3])", }, ) self.validate_all( "has([1], x)", read={ "postgres": "x = any(array[1])", }, ) self.validate_all( "NOT has([1], x)", read={ "postgres": "any(array[1]) <> x", }, ) self.validate_all( "SELECT CAST('2020-01-01' AS TIMESTAMP) + INTERVAL '500' MICROSECOND", read={ "duckdb": "SELECT TIMESTAMP '2020-01-01' + INTERVAL '500 us'", "postgres": "SELECT TIMESTAMP '2020-01-01' + INTERVAL '500 us'", }, ) self.validate_all( "SELECT CURRENT_DATE()", read={ "clickhouse": "SELECT CURRENT_DATE()", "postgres": "SELECT CURRENT_DATE", }, ) self.validate_all( "SELECT CURRENT_TIMESTAMP()", read={ "clickhouse": "SELECT CURRENT_TIMESTAMP()", "postgres": "SELECT CURRENT_TIMESTAMP", }, ) self.validate_all( "SELECT match('ThOmAs', CONCAT('(?i)', 'thomas'))", read={ "postgres": "SELECT 'ThOmAs' ~* 'thomas'", }, ) self.validate_all( "SELECT match('ThOmAs', CONCAT('(?i)', x)) FROM t", read={ "postgres": "SELECT 'ThOmAs' ~* x FROM t", }, ) self.validate_all( "SELECT '\\0'", read={ "mysql": "SELECT '\0'", }, write={ "clickhouse": "SELECT '\\0'", "mysql": "SELECT '\0'", }, ) self.validate_all( "DATE_ADD(DAY, 1, x)", read={ "clickhouse": "dateAdd(DAY, 1, x)", "presto": "DATE_ADD('DAY', 1, x)", }, write={ "clickhouse": "DATE_ADD(DAY, 1, x)", "presto": "DATE_ADD('DAY', 1, x)", "": "DATE_ADD(x, 1, 'DAY')", }, ) self.validate_all( "DATE_DIFF(DAY, a, b)", read={ "clickhouse": "dateDiff(DAY, a, b)", "presto": "DATE_DIFF('DAY', a, b)", }, write={ "clickhouse": "DATE_DIFF(DAY, a, b)", "presto": "DATE_DIFF('DAY', a, b)", "": "DATEDIFF(b, a, DAY)", }, ) self.validate_all( "SELECT xor(1, 0)", read={ "clickhouse": "SELECT xor(1, 0)", "mysql": "SELECT 1 XOR 0", }, write={ "mysql": "SELECT 1 XOR 0", }, ) self.validate_all( "SELECT xor(0, 1, xor(1, 0, 0))", write={ "clickhouse": "SELECT xor(0, 1, xor(1, 0, 0))", "mysql": "SELECT 0 XOR 1 XOR 1 XOR 0 XOR 0", }, ) self.validate_all( "SELECT xor(xor(1, 0), 1)", read={ "clickhouse": "SELECT xor(xor(1, 0), 1)", "mysql": "SELECT 1 XOR 0 XOR 1", }, write={ "clickhouse": "SELECT xor(xor(1, 0), 1)", "mysql": "SELECT 1 XOR 0 XOR 1", }, ) self.validate_all( "CONCAT(a, b)", read={ "clickhouse": "CONCAT(a, b)", "mysql": "CONCAT(a, b)", }, write={ "mysql": "CONCAT(a, b)", "postgres": "CONCAT(a, b)", }, ) self.validate_all( r"'Enum8(\'Sunday\' = 0)'", write={"clickhouse": "'Enum8(''Sunday'' = 0)'"} ) self.validate_all( "SELECT uniq(x) FROM (SELECT any(y) AS x FROM (SELECT 1 AS y))", read={ "bigquery": "SELECT APPROX_COUNT_DISTINCT(x) FROM (SELECT ANY_VALUE(y) x FROM (SELECT 1 y))", }, write={ "bigquery": "SELECT APPROX_COUNT_DISTINCT(x) FROM (SELECT ANY_VALUE(y) AS x FROM (SELECT 1 AS y))", }, ) self.validate_all( "SELECT fname, lname, age FROM person ORDER BY age DESC NULLS FIRST, fname ASC NULLS LAST, lname", write={ "clickhouse": "SELECT fname, lname, age FROM person ORDER BY age DESC NULLS FIRST, fname ASC, lname", "spark": "SELECT fname, lname, age FROM person ORDER BY age DESC NULLS FIRST, fname ASC NULLS LAST, lname NULLS LAST", }, ) self.validate_all( "CAST(1 AS NULLABLE(Int64))", write={ "clickhouse": "CAST(1 AS Nullable(Int64))", }, ) self.validate_all( "CAST(1 AS Nullable(DateTime64(6, 'UTC')))", write={"clickhouse": "CAST(1 AS Nullable(DateTime64(6, 'UTC')))"}, ) self.validate_all( "SELECT x #! comment", write={"": "SELECT x /* comment */"}, ) self.validate_all( "SELECT quantileIf(0.5)(a, true)", write={ "clickhouse": "SELECT quantileIf(0.5)(a, TRUE)", }, ) self.validate_all( "SELECT position(needle IN haystack)", write={"clickhouse": "SELECT position(haystack, needle)"}, ) self.validate_identity( "SELECT * FROM x LIMIT 10 SETTINGS max_results = 100, result = 'break'" ) self.validate_identity("SELECT * FROM x LIMIT 10 SETTINGS max_results = 100, result_") self.validate_identity("SELECT * FROM x FORMAT PrettyCompact") self.validate_identity( "SELECT * FROM x LIMIT 10 SETTINGS max_results = 100, result_ FORMAT PrettyCompact" ) self.validate_all( "SELECT * FROM foo JOIN bar USING id, name", write={"clickhouse": "SELECT * FROM foo JOIN bar USING (id, name)"}, ) self.validate_all( "SELECT * FROM foo ANY LEFT JOIN bla ON foo.c1 = bla.c2", write={"clickhouse": "SELECT * FROM foo LEFT ANY JOIN bla ON foo.c1 = bla.c2"}, ) self.validate_all( "SELECT * FROM foo GLOBAL ANY LEFT JOIN bla ON foo.c1 = bla.c2", write={"clickhouse": "SELECT * FROM foo GLOBAL LEFT ANY JOIN bla ON foo.c1 = bla.c2"}, ) self.validate_all( """ SELECT loyalty, count() FROM hits SEMI LEFT JOIN users USING (UserID) GROUP BY loyalty ORDER BY loyalty ASC """, write={ "clickhouse": "SELECT loyalty, count() FROM hits LEFT SEMI JOIN users USING (UserID)" " GROUP BY loyalty ORDER BY loyalty ASC" }, ) self.validate_identity("SELECT s, arr FROM arrays_test ARRAY JOIN arr") self.validate_identity("SELECT s, arr, a FROM arrays_test LEFT ARRAY JOIN arr AS a") self.validate_identity( "SELECT s, arr_external FROM arrays_test ARRAY JOIN [1, 2, 3] AS arr_external" ) self.validate_all( "SELECT quantile(0.5)(a)", read={"duckdb": "SELECT quantile(a, 0.5)"}, write={"clickhouse": "SELECT quantile(0.5)(a)"}, ) self.validate_all( "SELECT quantiles(0.5, 0.4)(a)", read={"duckdb": "SELECT quantile(a, [0.5, 0.4])"}, write={"clickhouse": "SELECT quantiles(0.5, 0.4)(a)"}, ) self.validate_all( "SELECT quantiles(0.5)(a)", read={"duckdb": "SELECT quantile(a, [0.5])"}, write={"clickhouse": "SELECT quantiles(0.5)(a)"}, ) self.validate_identity("SELECT isNaN(x)") self.validate_all( "SELECT IS_NAN(x), ISNAN(x)", write={"clickhouse": "SELECT isNaN(x), isNaN(x)"}, ) self.validate_identity("SELECT startsWith('a', 'b')") self.validate_all( "SELECT STARTS_WITH('a', 'b'), STARTSWITH('a', 'b')", write={"clickhouse": "SELECT startsWith('a', 'b'), startsWith('a', 'b')"}, ) def test_cte(self): self.validate_identity("WITH 'x' AS foo SELECT foo") self.validate_identity("WITH SUM(bytes) AS foo SELECT foo FROM system.parts") self.validate_identity("WITH (SELECT foo) AS bar SELECT bar + 5") self.validate_identity("WITH test1 AS (SELECT i + 1, j + 1 FROM test1) SELECT * FROM test1") query = parse_one("""WITH (SELECT 1) AS y SELECT * FROM y""", read="clickhouse") self.assertIsInstance(query.args["with"].expressions[0].this, exp.Subquery) self.assertEqual(query.args["with"].expressions[0].alias, "y") def test_ternary(self): self.validate_all("x ? 1 : 2", write={"clickhouse": "CASE WHEN x THEN 1 ELSE 2 END"}) self.validate_all( "IF(BAR(col), sign > 0 ? FOO() : 0, 1)", write={ "clickhouse": "CASE WHEN BAR(col) THEN CASE WHEN sign > 0 THEN FOO() ELSE 0 END ELSE 1 END" }, ) self.validate_all( "x AND FOO() > 3 + 2 ? 1 : 2", write={"clickhouse": "CASE WHEN x AND FOO() > 3 + 2 THEN 1 ELSE 2 END"}, ) self.validate_all( "x ? (y ? 1 : 2) : 3", write={"clickhouse": "CASE WHEN x THEN (CASE WHEN y THEN 1 ELSE 2 END) ELSE 3 END"}, ) self.validate_all( "x AND (foo() ? FALSE : TRUE) ? (y ? 1 : 2) : 3", write={ "clickhouse": "CASE WHEN x AND (CASE WHEN foo() THEN FALSE ELSE TRUE END) THEN (CASE WHEN y THEN 1 ELSE 2 END) ELSE 3 END" }, ) ternary = parse_one("x ? (y ? 1 : 2) : 3", read="clickhouse") self.assertIsInstance(ternary, exp.If) self.assertIsInstance(ternary.this, exp.Column) self.assertIsInstance(ternary.args["true"], exp.Paren) self.assertIsInstance(ternary.args["false"], exp.Literal) nested_ternary = ternary.args["true"].this self.assertIsInstance(nested_ternary.this, exp.Column) self.assertIsInstance(nested_ternary.args["true"], exp.Literal) self.assertIsInstance(nested_ternary.args["false"], exp.Literal) parse_one("a and b ? 1 : 2", read="clickhouse").assert_is(exp.If).this.assert_is(exp.And) def test_parameterization(self): self.validate_all( "SELECT {abc: UInt32}, {b: String}, {c: DateTime},{d: Map(String, Array(UInt8))}, {e: Tuple(UInt8, String)}", write={ "clickhouse": "SELECT {abc: UInt32}, {b: String}, {c: DATETIME}, {d: Map(String, Array(UInt8))}, {e: Tuple(UInt8, String)}", "": "SELECT :abc, :b, :c, :d, :e", }, ) self.validate_all( "SELECT * FROM {table: Identifier}", write={"clickhouse": "SELECT * FROM {table: Identifier}"}, ) def test_signed_and_unsigned_types(self): data_types = [ "UInt8", "UInt16", "UInt32", "UInt64", "UInt128", "UInt256", "Int8", "Int16", "Int32", "Int64", "Int128", "Int256", ] for data_type in data_types: self.validate_all( f"pow(2, 32)::{data_type}", write={"clickhouse": f"CAST(pow(2, 32) AS {data_type})"}, ) def test_ddl(self): self.validate_identity( 'CREATE TABLE data5 ("x" UInt32, "y" UInt32) ENGINE=MergeTree ORDER BY (round(y / 1000000000), cityHash64(x)) SAMPLE BY cityHash64(x)' ) self.validate_identity( "CREATE TABLE foo (x UInt32) TTL time_column + INTERVAL '1' MONTH DELETE WHERE column = 'value'" ) self.validate_all( """ CREATE TABLE example1 ( timestamp DateTime, x UInt32 TTL now() + INTERVAL 1 MONTH, y String TTL timestamp + INTERVAL 1 DAY, z String ) ENGINE = MergeTree ORDER BY tuple() """, write={ "clickhouse": """CREATE TABLE example1 ( timestamp DATETIME, x UInt32 TTL now() + INTERVAL '1' MONTH, y String TTL timestamp + INTERVAL '1' DAY, z String ) ENGINE=MergeTree ORDER BY tuple()""", }, pretty=True, ) self.validate_all( """ CREATE TABLE test (id UInt64, timestamp DateTime64, data String, max_hits UInt64, sum_hits UInt64) ENGINE = MergeTree PRIMARY KEY (id, toStartOfDay(timestamp), timestamp) TTL timestamp + INTERVAL 1 DAY GROUP BY id, toStartOfDay(timestamp) SET max_hits = max(max_hits), sum_hits = sum(sum_hits) """, write={ "clickhouse": """CREATE TABLE test ( id UInt64, timestamp DateTime64, data String, max_hits UInt64, sum_hits UInt64 ) ENGINE=MergeTree PRIMARY KEY (id, toStartOfDay(timestamp), timestamp) TTL timestamp + INTERVAL '1' DAY GROUP BY id, toStartOfDay(timestamp) SET max_hits = max(max_hits), sum_hits = sum(sum_hits)""", }, pretty=True, ) self.validate_all( """ CREATE TABLE test (id String, data String) ENGINE = AggregatingMergeTree() ORDER BY tuple() SETTINGS max_suspicious_broken_parts=500, parts_to_throw_insert=100 """, write={ "clickhouse": """CREATE TABLE test ( id String, data String ) ENGINE=AggregatingMergeTree() ORDER BY tuple() SETTINGS max_suspicious_broken_parts = 500, parts_to_throw_insert = 100""", }, pretty=True, ) self.validate_all( """ CREATE TABLE example_table ( d DateTime, a Int ) ENGINE = MergeTree PARTITION BY toYYYYMM(d) ORDER BY d TTL d + INTERVAL 1 MONTH DELETE, d + INTERVAL 1 WEEK TO VOLUME 'aaa', d + INTERVAL 2 WEEK TO DISK 'bbb'; """, write={ "clickhouse": """CREATE TABLE example_table ( d DATETIME, a Int32 ) ENGINE=MergeTree PARTITION BY toYYYYMM(d) ORDER BY d TTL d + INTERVAL '1' MONTH DELETE, d + INTERVAL '1' WEEK TO VOLUME 'aaa', d + INTERVAL '2' WEEK TO DISK 'bbb'""", }, pretty=True, ) self.validate_all( """ CREATE TABLE table_with_where ( d DateTime, a Int ) ENGINE = MergeTree PARTITION BY toYYYYMM(d) ORDER BY d TTL d + INTERVAL 1 MONTH DELETE WHERE toDayOfWeek(d) = 1; """, write={ "clickhouse": """CREATE TABLE table_with_where ( d DATETIME, a Int32 ) ENGINE=MergeTree PARTITION BY toYYYYMM(d) ORDER BY d TTL d + INTERVAL '1' MONTH DELETE WHERE toDayOfWeek(d) = 1""", }, pretty=True, ) self.validate_all( """ CREATE TABLE table_for_recompression ( d DateTime, key UInt64, value String ) ENGINE MergeTree() ORDER BY tuple() PARTITION BY key TTL d + INTERVAL 1 MONTH RECOMPRESS CODEC(ZSTD(17)), d + INTERVAL 1 YEAR RECOMPRESS CODEC(LZ4HC(10)) SETTINGS min_rows_for_wide_part = 0, min_bytes_for_wide_part = 0; """, write={ "clickhouse": """CREATE TABLE table_for_recompression ( d DATETIME, key UInt64, value String ) ENGINE=MergeTree() ORDER BY tuple() PARTITION BY key TTL d + INTERVAL '1' MONTH RECOMPRESS CODEC(ZSTD(17)), d + INTERVAL '1' YEAR RECOMPRESS CODEC(LZ4HC(10)) SETTINGS min_rows_for_wide_part = 0, min_bytes_for_wide_part = 0""", }, pretty=True, ) self.validate_all( """ CREATE TABLE table_for_aggregation ( d DateTime, k1 Int, k2 Int, x Int, y Int ) ENGINE = MergeTree ORDER BY (k1, k2) TTL d + INTERVAL 1 MONTH GROUP BY k1, k2 SET x = max(x), y = min(y); """, write={ "clickhouse": """CREATE TABLE table_for_aggregation ( d DATETIME, k1 Int32, k2 Int32, x Int32, y Int32 ) ENGINE=MergeTree ORDER BY (k1, k2) TTL d + INTERVAL '1' MONTH GROUP BY k1, k2 SET x = max(x), y = min(y)""", }, pretty=True, ) self.validate_all( """ CREATE DICTIONARY discounts_dict ( advertiser_id UInt64, discount_start_date Date, discount_end_date Date, amount Float64 ) PRIMARY KEY id SOURCE(CLICKHOUSE(TABLE 'discounts')) LIFETIME(MIN 1 MAX 1000) LAYOUT(RANGE_HASHED(range_lookup_strategy 'max')) RANGE(MIN discount_start_date MAX discount_end_date) """, write={ "clickhouse": """CREATE DICTIONARY discounts_dict ( advertiser_id UInt64, discount_start_date DATE, discount_end_date DATE, amount Float64 ) PRIMARY KEY (id) SOURCE(CLICKHOUSE( TABLE 'discounts' )) LIFETIME(MIN 1 MAX 1000) LAYOUT(RANGE_HASHED( range_lookup_strategy 'max' )) RANGE(MIN discount_start_date MAX discount_end_date)""", }, pretty=True, ) self.validate_all( """ CREATE DICTIONARY my_ip_trie_dictionary ( prefix String, asn UInt32, cca2 String DEFAULT '??' ) PRIMARY KEY prefix SOURCE(CLICKHOUSE(TABLE 'my_ip_addresses')) LAYOUT(IP_TRIE) LIFETIME(3600); """, write={ "clickhouse": """CREATE DICTIONARY my_ip_trie_dictionary ( prefix String, asn UInt32, cca2 String DEFAULT '??' ) PRIMARY KEY (prefix) SOURCE(CLICKHOUSE( TABLE 'my_ip_addresses' )) LAYOUT(IP_TRIE()) LIFETIME(MIN 0 MAX 3600)""", }, pretty=True, ) self.validate_all( """ CREATE DICTIONARY polygons_test_dictionary ( key Array(Array(Array(Tuple(Float64, Float64)))), name String ) PRIMARY KEY key SOURCE(CLICKHOUSE(TABLE 'polygons_test_table')) LAYOUT(POLYGON(STORE_POLYGON_KEY_COLUMN 1)) LIFETIME(0); """, write={ "clickhouse": """CREATE DICTIONARY polygons_test_dictionary ( key Array(Array(Array(Tuple(Float64, Float64)))), name String ) PRIMARY KEY (key) SOURCE(CLICKHOUSE( TABLE 'polygons_test_table' )) LAYOUT(POLYGON( STORE_POLYGON_KEY_COLUMN 1 )) LIFETIME(MIN 0 MAX 0)""", }, pretty=True, ) self.validate_identity("""CREATE TABLE ip_data (ip4 IPv4, ip6 IPv6) ENGINE=TinyLog()""") self.validate_identity("""CREATE TABLE dates (dt1 Date32) ENGINE=TinyLog()""") self.validate_all( """ CREATE TABLE t ( a AggregateFunction(quantiles(0.5, 0.9), UInt64), b AggregateFunction(quantiles, UInt64), c SimpleAggregateFunction(sum, Float64) )""", write={ "clickhouse": """CREATE TABLE t ( a AggregateFunction(quantiles(0.5, 0.9), UInt64), b AggregateFunction(quantiles, UInt64), c SimpleAggregateFunction(sum, Float64) )""" }, pretty=True, )