fix(tsql, postgres)!: Improve UUID support (#4718)

* fix(tsql, postgres): Improve UUID support

* Move NEWID reference into T-SQL specific dialect

Author: amachanic

sqlglot/dialects/spark.py

@@ -144,7 +144,7 @@ class Generator(Spark2.Generator):
             **Spark2.Generator.TYPE_MAPPING,
             exp.DataType.Type.MONEY: "DECIMAL(15, 4)",
             exp.DataType.Type.SMALLMONEY: "DECIMAL(6, 4)",
-            exp.DataType.Type.UNIQUEIDENTIFIER: "STRING",
+            exp.DataType.Type.UUID: "STRING",
             exp.DataType.Type.TIMESTAMPLTZ: "TIMESTAMP_LTZ",
             exp.DataType.Type.TIMESTAMPNTZ: "TIMESTAMP_NTZ",
         }

sqlglot/dialects/tsql.py

@@ -525,7 +525,7 @@ class Tokenizer(tokens.Tokenizer):
             "TOP": TokenType.TOP,
             "TIMESTAMP": TokenType.ROWVERSION,
             "TINYINT": TokenType.UTINYINT,
-            "UNIQUEIDENTIFIER": TokenType.UNIQUEIDENTIFIER,
+            "UNIQUEIDENTIFIER": TokenType.UUID,
             "UPDATE STATISTICS": TokenType.COMMAND,
             "XML": TokenType.XML,
         }
@@ -579,6 +579,7 @@ class Parser(parser.Parser):
             "JSON_VALUE": parser.build_extract_json_with_path(exp.JSONExtractScalar),
             "LEN": _build_with_arg_as_text(exp.Length),
             "LEFT": _build_with_arg_as_text(exp.Left),
+            "NEWID": exp.Uuid.from_arg_list,
             "RIGHT": _build_with_arg_as_text(exp.Right),
             "PARSENAME": _build_parsename,
             "REPLICATE": exp.Repeat.from_arg_list,
@@ -919,6 +920,7 @@ class Generator(generator.Generator):
             exp.DataType.Type.SMALLDATETIME: "SMALLDATETIME",
             exp.DataType.Type.UTINYINT: "TINYINT",
             exp.DataType.Type.VARIANT: "SQL_VARIANT",
+            exp.DataType.Type.UUID: "UNIQUEIDENTIFIER",
         }
 
         TYPE_MAPPING.pop(exp.DataType.Type.NCHAR)
@@ -974,6 +976,7 @@ class Generator(generator.Generator):
             exp.TsOrDsAdd: date_delta_sql("DATEADD", cast=True),
             exp.TsOrDsDiff: date_delta_sql("DATEDIFF"),
             exp.TimestampTrunc: lambda self, e: self.func("DATETRUNC", e.unit, e.this),
+            exp.Uuid: lambda *_: "NEWID()",
             exp.DateFromParts: rename_func("DATEFROMPARTS"),
         }
 

sqlglot/expressions.py

@@ -4476,7 +4476,6 @@ class Type(AutoName):
         UMEDIUMINT = auto()
         UDECIMAL = auto()
         UNION = auto()
-        UNIQUEIDENTIFIER = auto()
         UNKNOWN = auto()  # Sentinel value, useful for type annotation
         USERDEFINED = "USER-DEFINED"
         USMALLINT = auto()

sqlglot/parser.py

@@ -387,7 +387,6 @@ class Parser(metaclass=_Parser):
         TokenType.BIGSERIAL,
         TokenType.XML,
         TokenType.YEAR,
-        TokenType.UNIQUEIDENTIFIER,
         TokenType.USERDEFINED,
         TokenType.MONEY,
         TokenType.SMALLMONEY,

sqlglot/tokens.py

@@ -192,7 +192,6 @@ class TokenType(AutoName):
     BIGSERIAL = auto()
     XML = auto()
     YEAR = auto()
-    UNIQUEIDENTIFIER = auto()
     USERDEFINED = auto()
     MONEY = auto()
     SMALLMONEY = auto()

tests/dialects/test_dialect.py

@@ -3184,6 +3184,7 @@ def test_uuid(self):
                 "postgres": "GEN_RANDOM_UUID()",
                 "bigquery": "GENERATE_UUID()",
                 "snowflake": "UUID_STRING()",
+                "tsql": "NEWID()",
             },
             write={
                 "hive": "UUID()",
@@ -3197,6 +3198,7 @@ def test_uuid(self):
                 "postgres": "GEN_RANDOM_UUID()",
                 "bigquery": "GENERATE_UUID()",
                 "snowflake": "UUID_STRING()",
+                "tsql": "NEWID()",
             },
         )
 

tests/dialects/test_postgres.py

@@ -1088,7 +1088,8 @@ def test_ddl(self):
                 "duckdb": "CREATE TABLE x (a UUID, b BLOB)",
                 "presto": "CREATE TABLE x (a UUID, b VARBINARY)",
                 "hive": "CREATE TABLE x (a UUID, b BINARY)",
-                "spark": "CREATE TABLE x (a UUID, b BINARY)",
+                "spark": "CREATE TABLE x (a STRING, b BINARY)",
+                "tsql": "CREATE TABLE x (a UNIQUEIDENTIFIER, b VARBINARY)",
             },
         )
 

tests/dialects/test_tsql.py

@@ -928,6 +928,15 @@ def test_ddl(self):
                 "databricks": "CREATE TABLE tbl (id BIGINT NOT NULL GENERATED ALWAYS AS IDENTITY (START WITH 10 INCREMENT BY 1) PRIMARY KEY)",
             },
         )
+        self.validate_all(
+            "CREATE TABLE x (a UNIQUEIDENTIFIER, b VARBINARY)",
+            write={
+                "duckdb": "CREATE TABLE x (a UUID, b BLOB)",
+                "presto": "CREATE TABLE x (a UUID, b VARBINARY)",
+                "spark": "CREATE TABLE x (a STRING, b BINARY)",
+                "postgres": "CREATE TABLE x (a UUID, b BYTEA)",
+            },
+        )
         self.validate_all(
             "SELECT * INTO foo.bar.baz FROM (SELECT * FROM a.b.c) AS temp",
             read={