from sqlglot import ErrorLevel, ParseError, UnsupportedError, transpile from tests.dialects.test_dialect import Validator class TestBigQuery(Validator): dialect = "bigquery" def test_bigquery(self): self.validate_all( "REGEXP_CONTAINS('foo', '.*')", read={"bigquery": "REGEXP_CONTAINS('foo', '.*')"}, write={"mysql": "REGEXP_LIKE('foo', '.*')"}, ), self.validate_all( '"""x"""', write={ "bigquery": "'x'", "duckdb": "'x'", "presto": "'x'", "hive": "'x'", "spark": "'x'", }, ) self.validate_all( '"""x\'"""', write={ "bigquery": "'x\\''", "duckdb": "'x'''", "presto": "'x'''", "hive": "'x\\''", "spark": "'x\\''", }, ) self.validate_all( r'r"""/\*.*\*/"""', write={ "bigquery": r"'/\\*.*\\*/'", "duckdb": r"'/\*.*\*/'", "presto": r"'/\*.*\*/'", "hive": r"'/\\*.*\\*/'", "spark": r"'/\\*.*\\*/'", }, ) self.validate_all( R'R"""/\*.*\*/"""', write={ "bigquery": R"'/\\*.*\\*/'", "duckdb": R"'/\*.*\*/'", "presto": R"'/\*.*\*/'", "hive": R"'/\\*.*\\*/'", "spark": R"'/\\*.*\\*/'", }, ) self.validate_all( "CAST(a AS INT64)", write={ "bigquery": "CAST(a AS INT64)", "duckdb": "CAST(a AS BIGINT)", "presto": "CAST(a AS BIGINT)", "hive": "CAST(a AS BIGINT)", "spark": "CAST(a AS LONG)", }, ) self.validate_all( "CAST(a AS NUMERIC)", write={ "bigquery": "CAST(a AS NUMERIC)", "duckdb": "CAST(a AS DECIMAL)", "presto": "CAST(a AS DECIMAL)", "hive": "CAST(a AS DECIMAL)", "spark": "CAST(a AS DECIMAL)", }, ) self.validate_all( "[1, 2, 3]", read={ "duckdb": "LIST_VALUE(1, 2, 3)", "presto": "ARRAY[1, 2, 3]", "hive": "ARRAY(1, 2, 3)", "spark": "ARRAY(1, 2, 3)", }, write={ "bigquery": "[1, 2, 3]", "duckdb": "LIST_VALUE(1, 2, 3)", "presto": "ARRAY[1, 2, 3]", "hive": "ARRAY(1, 2, 3)", "spark": "ARRAY(1, 2, 3)", }, ) self.validate_all( "SELECT * FROM UNNEST(['7', '14']) AS x", read={ "spark": "SELECT * FROM UNNEST(ARRAY('7', '14')) AS (x)", }, write={ "bigquery": "SELECT * FROM UNNEST(['7', '14']) AS x", "presto": "SELECT * FROM UNNEST(ARRAY['7', '14']) AS (x)", "hive": "SELECT * FROM UNNEST(ARRAY('7', '14')) AS (x)", "spark": "SELECT * FROM UNNEST(ARRAY('7', '14')) AS (x)", }, ) self.validate_all( "SELECT ARRAY(SELECT x FROM UNNEST([0, 1]) AS x)", write={"bigquery": "SELECT ARRAY(SELECT x FROM UNNEST([0, 1]) AS x)"}, ) self.validate_all( "SELECT ARRAY(SELECT DISTINCT x FROM UNNEST(some_numbers) AS x) AS unique_numbers", write={ "bigquery": "SELECT ARRAY(SELECT DISTINCT x FROM UNNEST(some_numbers) AS x) AS unique_numbers" }, ) self.validate_all( "SELECT ARRAY(SELECT * FROM foo JOIN bla ON x = y)", write={"bigquery": "SELECT ARRAY(SELECT * FROM foo JOIN bla ON x = y)"}, ) self.validate_all( "x IS unknown", write={ "bigquery": "x IS NULL", "duckdb": "x IS NULL", "presto": "x IS NULL", "hive": "x IS NULL", "spark": "x IS NULL", }, ) self.validate_all( "CURRENT_DATE", read={ "tsql": "GETDATE()", }, write={ "tsql": "GETDATE()", }, ) self.validate_all( "current_datetime", write={ "bigquery": "CURRENT_DATETIME()", "duckdb": "CURRENT_DATETIME()", "presto": "CURRENT_DATETIME()", "hive": "CURRENT_DATETIME()", "spark": "CURRENT_DATETIME()", }, ) self.validate_all( "current_time", write={ "bigquery": "CURRENT_TIME()", "duckdb": "CURRENT_TIME()", "presto": "CURRENT_TIME()", "hive": "CURRENT_TIME()", "spark": "CURRENT_TIME()", }, ) self.validate_all( "current_timestamp", write={ "bigquery": "CURRENT_TIMESTAMP()", "duckdb": "CURRENT_TIMESTAMP()", "postgres": "CURRENT_TIMESTAMP", "presto": "CURRENT_TIMESTAMP()", "hive": "CURRENT_TIMESTAMP()", "spark": "CURRENT_TIMESTAMP()", }, ) self.validate_all( "current_timestamp()", write={ "bigquery": "CURRENT_TIMESTAMP()", "duckdb": "CURRENT_TIMESTAMP()", "postgres": "CURRENT_TIMESTAMP", "presto": "CURRENT_TIMESTAMP()", "hive": "CURRENT_TIMESTAMP()", "spark": "CURRENT_TIMESTAMP()", }, ) self.validate_all( "DIV(x, y)", write={ "bigquery": "DIV(x, y)", "duckdb": "CAST(x / y AS INT)", }, ) self.validate_identity( "SELECT ROW() OVER (y ROWS BETWEEN 1 PRECEDING AND 1 FOLLOWING) FROM x WINDOW y AS (PARTITION BY CATEGORY)" ) self.validate_identity( "SELECT item, purchases, LAST_VALUE(item) OVER (item_window ROWS BETWEEN 2 PRECEDING AND 2 FOLLOWING) AS most_popular FROM Produce WINDOW item_window AS (ORDER BY purchases)" ) self.validate_identity( "SELECT LAST_VALUE(a IGNORE NULLS) OVER y FROM x WINDOW y AS (PARTITION BY CATEGORY)", ) self.validate_all( "CREATE TABLE db.example_table (col_a struct)", write={ "bigquery": "CREATE TABLE db.example_table (col_a STRUCT)", "duckdb": "CREATE TABLE db.example_table (col_a STRUCT(struct_col_a INT, struct_col_b TEXT))", "presto": "CREATE TABLE db.example_table (col_a ROW(struct_col_a INTEGER, struct_col_b VARCHAR))", "hive": "CREATE TABLE db.example_table (col_a STRUCT)", "spark": "CREATE TABLE db.example_table (col_a STRUCT)", }, ) self.validate_all( "CREATE TABLE db.example_table (col_a STRUCT>)", write={ "bigquery": "CREATE TABLE db.example_table (col_a STRUCT>)", "duckdb": "CREATE TABLE db.example_table (col_a STRUCT(struct_col_a BIGINT, struct_col_b STRUCT(nested_col_a TEXT, nested_col_b TEXT)))", "presto": "CREATE TABLE db.example_table (col_a ROW(struct_col_a BIGINT, struct_col_b ROW(nested_col_a VARCHAR, nested_col_b VARCHAR)))", "hive": "CREATE TABLE db.example_table (col_a STRUCT>)", "spark": "CREATE TABLE db.example_table (col_a STRUCT>)", }, ) self.validate_all( "SELECT * FROM a WHERE b IN UNNEST([1, 2, 3])", write={ "bigquery": "SELECT * FROM a WHERE b IN UNNEST([1, 2, 3])", "mysql": "SELECT * FROM a WHERE b IN (SELECT UNNEST(ARRAY(1, 2, 3)))", "presto": "SELECT * FROM a WHERE b IN (SELECT UNNEST(ARRAY[1, 2, 3]))", "hive": "SELECT * FROM a WHERE b IN (SELECT UNNEST(ARRAY(1, 2, 3)))", "spark": "SELECT * FROM a WHERE b IN (SELECT UNNEST(ARRAY(1, 2, 3)))", }, ) # Reference: https://cloud.google.com/bigquery/docs/reference/standard-sql/query-syntax#set_operators with self.assertRaises(UnsupportedError): transpile( "SELECT * FROM a INTERSECT ALL SELECT * FROM b", write="bigquery", unsupported_level=ErrorLevel.RAISE, ) with self.assertRaises(UnsupportedError): transpile( "SELECT * FROM a EXCEPT ALL SELECT * FROM b", write="bigquery", unsupported_level=ErrorLevel.RAISE, ) with self.assertRaises(ParseError): transpile("SELECT * FROM UNNEST(x) AS x(y)", read="bigquery") self.validate_all( "DATE_SUB(CURRENT_DATE(), INTERVAL 1 DAY)", write={ "postgres": "CURRENT_DATE - INTERVAL '1' DAY", }, ) self.validate_all( "DATE_ADD(CURRENT_DATE(), INTERVAL 1 DAY)", write={ "bigquery": "DATE_ADD(CURRENT_DATE, INTERVAL 1 DAY)", "duckdb": "CURRENT_DATE + INTERVAL 1 DAY", "mysql": "DATE_ADD(CURRENT_DATE, INTERVAL 1 DAY)", "postgres": "CURRENT_DATE + INTERVAL '1' DAY", "presto": "DATE_ADD(DAY, 1, CURRENT_DATE)", "hive": "DATE_ADD(CURRENT_DATE, 1)", "spark": "DATE_ADD(CURRENT_DATE, 1)", }, ) self.validate_all( "DATE_DIFF(DATE '2010-07-07', DATE '2008-12-25', DAY)", write={ "bigquery": "DATE_DIFF(CAST('2010-07-07' AS DATE), CAST('2008-12-25' AS DATE), DAY)", "mysql": "DATEDIFF(CAST('2010-07-07' AS DATE), CAST('2008-12-25' AS DATE))", }, ) self.validate_all( "DATE_DIFF(DATE '2010-07-07', DATE '2008-12-25', MINUTE)", write={ "bigquery": "DATE_DIFF(CAST('2010-07-07' AS DATE), CAST('2008-12-25' AS DATE), MINUTE)", "mysql": "DATEDIFF(CAST('2010-07-07' AS DATE), CAST('2008-12-25' AS DATE))", }, ) self.validate_all( "CURRENT_DATE('UTC')", write={ "mysql": "CURRENT_DATE AT TIME ZONE 'UTC'", "postgres": "CURRENT_DATE AT TIME ZONE 'UTC'", }, ) self.validate_all( "SELECT a FROM test WHERE a = 1 GROUP BY a HAVING a = 2 QUALIFY z ORDER BY a LIMIT 10", write={ "bigquery": "SELECT a FROM test WHERE a = 1 GROUP BY a HAVING a = 2 QUALIFY z ORDER BY a LIMIT 10", "snowflake": "SELECT a FROM test WHERE a = 1 GROUP BY a HAVING a = 2 QUALIFY z ORDER BY a NULLS FIRST LIMIT 10", }, ) self.validate_all( "SELECT cola, colb FROM (VALUES (1, 'test')) AS tab(cola, colb)", write={ "spark": "SELECT cola, colb FROM VALUES (1, 'test') AS tab(cola, colb)", "bigquery": "SELECT cola, colb FROM UNNEST([STRUCT(1 AS cola, 'test' AS colb)])", "snowflake": "SELECT cola, colb FROM (VALUES (1, 'test')) AS tab(cola, colb)", }, ) self.validate_all( "SELECT * FROM (SELECT a, b, c FROM test) PIVOT(SUM(b) d, COUNT(*) e FOR c IN ('x', 'y'))", write={ "bigquery": "SELECT * FROM (SELECT a, b, c FROM test) PIVOT(SUM(b) AS d, COUNT(*) AS e FOR c IN ('x', 'y'))", }, ) self.validate_identity("BEGIN A B C D E F") self.validate_identity("BEGIN TRANSACTION") self.validate_identity("COMMIT TRANSACTION") self.validate_identity("ROLLBACK TRANSACTION") def test_user_defined_functions(self): self.validate_identity( "CREATE TEMPORARY FUNCTION a(x FLOAT64, y FLOAT64) RETURNS FLOAT64 NOT DETERMINISTIC LANGUAGE js AS 'return x*y;'" ) self.validate_identity("CREATE TEMPORARY FUNCTION a(x FLOAT64, y FLOAT64) AS ((x + 4) / y)") self.validate_identity( "CREATE TABLE FUNCTION a(x INT64) RETURNS TABLE AS SELECT s, t" )