Merging upstream version 6.2.0.

Signed-off-by: Daniel Baumann <daniel@debian.org>

sqlglot/dialects/__init__.py

@@ -14,3 +14,4 @@ from sqlglot.dialects.sqlite import SQLite
 from sqlglot.dialects.starrocks import StarRocks
 from sqlglot.dialects.tableau import Tableau
 from sqlglot.dialects.trino import Trino
+from sqlglot.dialects.tsql import TSQL

sqlglot/dialects/dialect.py

@@ -27,6 +27,7 @@ class Dialects(str, Enum):
     STARROCKS = "starrocks"
     TABLEAU = "tableau"
     TRINO = "trino"
+    TSQL = "tsql"
 
 
 class _Dialect(type):
@@ -53,7 +54,6 @@ class _Dialect(type):
         klass.parser_class = getattr(klass, "Parser", Parser)
         klass.generator_class = getattr(klass, "Generator", Generator)
 
-        klass.tokenizer = klass.tokenizer_class()
         klass.quote_start, klass.quote_end = list(klass.tokenizer_class._QUOTES.items())[0]
         klass.identifier_start, klass.identifier_end = list(klass.tokenizer_class._IDENTIFIERS.items())[0]
 
@@ -95,7 +95,6 @@ class Dialect(metaclass=_Dialect):
     tokenizer_class = None
     parser_class = None
     generator_class = None
-    tokenizer = None
 
     @classmethod
     def get_or_raise(cls, dialect):
@@ -138,6 +137,12 @@ class Dialect(metaclass=_Dialect):
     def transpile(self, code, **opts):
         return self.generate(self.parse(code), **opts)
 
+    @property
+    def tokenizer(self):
+        if not hasattr(self, "_tokenizer"):
+            self._tokenizer = self.tokenizer_class()
+        return self._tokenizer
+
     def parser(self, **opts):
         return self.parser_class(
             **{
@@ -170,7 +175,15 @@ class Dialect(metaclass=_Dialect):
 
 
 def rename_func(name):
-    return lambda self, expression: f"{name}({csv(*[self.sql(e) for e in expression.args.values()])})"
+    def _rename(self, expression):
+        args = (
+            self.expressions(expression, flat=True)
+            if isinstance(expression, exp.Func) and expression.is_var_len_args
+            else csv(*[self.sql(e) for e in expression.args.values()])
+        )
+        return f"{name}({args})"
+
+    return _rename
 
 
 def approx_count_distinct_sql(self, expression):

sqlglot/dialects/duckdb.py

@@ -108,7 +108,7 @@ class DuckDB(Dialect):
         TRANSFORMS = {
             **Generator.TRANSFORMS,
             exp.ApproxDistinct: approx_count_distinct_sql,
-            exp.Array: lambda self, e: f"LIST_VALUE({self.expressions(e, flat=True)})",
+            exp.Array: rename_func("LIST_VALUE"),
             exp.ArraySize: rename_func("ARRAY_LENGTH"),
             exp.ArraySort: _array_sort_sql,
             exp.ArraySum: rename_func("LIST_SUM"),

sqlglot/dialects/snowflake.py

@@ -106,6 +106,11 @@ class Snowflake(Dialect):
             "TO_TIMESTAMP": _snowflake_to_timestamp,
         }
 
+        FUNCTION_PARSERS = {
+            **Parser.FUNCTION_PARSERS,
+            "DATE_PART": lambda self: self._parse_extract(),
+        }
+
         COLUMN_OPERATORS = {
             **Parser.COLUMN_OPERATORS,
             TokenType.COLON: lambda self, this, path: self.expression(
@@ -118,10 +123,20 @@ class Snowflake(Dialect):
     class Tokenizer(Tokenizer):
         QUOTES = ["'", "$$"]
         ESCAPE = "\\"
+
+        SINGLE_TOKENS = {
+            **Tokenizer.SINGLE_TOKENS,
+            "$": TokenType.DOLLAR,  # needed to break for quotes
+        }
+
         KEYWORDS = {
             **Tokenizer.KEYWORDS,
             "QUALIFY": TokenType.QUALIFY,
             "DOUBLE PRECISION": TokenType.DOUBLE,
+            "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
+            "TIMESTAMP_NTZ": TokenType.TIMESTAMP,
+            "TIMESTAMP_TZ": TokenType.TIMESTAMPTZ,
+            "TIMESTAMPNTZ": TokenType.TIMESTAMP,
         }
 
     class Generator(Generator):
@@ -132,6 +147,11 @@ class Snowflake(Dialect):
             exp.UnixToTime: _unix_to_time,
         }
 
+        TYPE_MAPPING = {
+            **Generator.TYPE_MAPPING,
+            exp.DataType.Type.TIMESTAMP: "TIMESTAMPNTZ",
+        }
+
         def except_op(self, expression):
             if not expression.args.get("distinct", False):
                 self.unsupported("EXCEPT with All is not supported in Snowflake")

sqlglot/dialects/spark.py

@@ -82,6 +82,7 @@ class Spark(Hive):
 
         TRANSFORMS = {
             **{k: v for k, v in Hive.Generator.TRANSFORMS.items() if k not in {exp.ArraySort}},
+            exp.FileFormatProperty: lambda self, e: f"USING {e.text('value').upper()}",
             exp.ArraySum: lambda self, e: f"AGGREGATE({self.sql(e, 'this')}, 0, (acc, x) -> acc + x, acc -> acc)",
             exp.BitwiseLeftShift: rename_func("SHIFTLEFT"),
             exp.BitwiseRightShift: rename_func("SHIFTRIGHT"),

sqlglot/dialects/tsql.py

@@ -0,0 +1,38 @@
+from sqlglot import exp
+from sqlglot.dialects.dialect import Dialect
+from sqlglot.generator import Generator
+from sqlglot.tokens import Tokenizer, TokenType
+
+
+class TSQL(Dialect):
+    null_ordering = "nulls_are_small"
+    time_format = "'yyyy-mm-dd hh:mm:ss'"
+
+    class Tokenizer(Tokenizer):
+        IDENTIFIERS = ['"', ("[", "]")]
+
+        KEYWORDS = {
+            **Tokenizer.KEYWORDS,
+            "BIT": TokenType.BOOLEAN,
+            "REAL": TokenType.FLOAT,
+            "NTEXT": TokenType.TEXT,
+            "SMALLDATETIME": TokenType.DATETIME,
+            "DATETIMEOFFSET": TokenType.TIMESTAMPTZ,
+            "TIME": TokenType.TIMESTAMP,
+            "VARBINARY": TokenType.BINARY,
+            "IMAGE": TokenType.IMAGE,
+            "MONEY": TokenType.MONEY,
+            "SMALLMONEY": TokenType.SMALLMONEY,
+            "ROWVERSION": TokenType.ROWVERSION,
+            "SQL_VARIANT": TokenType.SQL_VARIANT,
+            "UNIQUEIDENTIFIER": TokenType.UNIQUEIDENTIFIER,
+            "XML": TokenType.XML,
+        }
+
+    class Generator(Generator):
+        TYPE_MAPPING = {
+            **Generator.TYPE_MAPPING,
+            exp.DataType.Type.BOOLEAN: "BIT",
+            exp.DataType.Type.INT: "INTEGER",
+            exp.DataType.Type.DECIMAL: "NUMERIC",
+        }