Merging upstream version 18.2.0.

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

sqlglot/dialects/mysql.py

@@ -8,6 +8,7 @@ from sqlglot.dialects.dialect import (
     arrow_json_extract_scalar_sql,
     datestrtodate_sql,
     format_time_lambda,
+    json_keyvalue_comma_sql,
     locate_to_strposition,
     max_or_greatest,
     min_or_least,
@@ -32,7 +33,7 @@ def _show_parser(*args: t.Any, **kwargs: t.Any) -> t.Callable[[MySQL.Parser], ex
     return _parse
 
 
-def _date_trunc_sql(self: generator.Generator, expression: exp.DateTrunc) -> str:
+def _date_trunc_sql(self: MySQL.Generator, expression: exp.DateTrunc) -> str:
     expr = self.sql(expression, "this")
     unit = expression.text("unit")
 
@@ -63,12 +64,12 @@ def _str_to_date(args: t.List) -> exp.StrToDate:
     return exp.StrToDate(this=seq_get(args, 0), format=date_format)
 
 
-def _str_to_date_sql(self: generator.Generator, expression: exp.StrToDate | exp.StrToTime) -> str:
+def _str_to_date_sql(self: MySQL.Generator, expression: exp.StrToDate | exp.StrToTime) -> str:
     date_format = self.format_time(expression)
     return f"STR_TO_DATE({self.sql(expression.this)}, {date_format})"
 
 
-def _trim_sql(self: generator.Generator, expression: exp.Trim) -> str:
+def _trim_sql(self: MySQL.Generator, expression: exp.Trim) -> str:
     target = self.sql(expression, "this")
     trim_type = self.sql(expression, "position")
     remove_chars = self.sql(expression, "expression")
@@ -83,8 +84,8 @@ def _trim_sql(self: generator.Generator, expression: exp.Trim) -> str:
     return f"TRIM({trim_type}{remove_chars}{from_part}{target})"
 
 
-def _date_add_sql(kind: str) -> t.Callable[[generator.Generator, exp.DateAdd | exp.DateSub], str]:
-    def func(self: generator.Generator, expression: exp.DateAdd | exp.DateSub) -> str:
+def _date_add_sql(kind: str) -> t.Callable[[MySQL.Generator, exp.DateAdd | exp.DateSub], str]:
+    def func(self: MySQL.Generator, expression: exp.DateAdd | exp.DateSub) -> str:
         this = self.sql(expression, "this")
         unit = expression.text("unit").upper() or "DAY"
         return f"DATE_{kind}({this}, {self.sql(exp.Interval(this=expression.expression.copy(), unit=unit))})"
@@ -93,6 +94,9 @@ def _date_add_sql(kind: str) -> t.Callable[[generator.Generator, exp.DateAdd | e
 
 
 class MySQL(Dialect):
+    # https://dev.mysql.com/doc/refman/8.0/en/identifiers.html
+    IDENTIFIERS_CAN_START_WITH_DIGIT = True
+
     TIME_FORMAT = "'%Y-%m-%d %T'"
     DPIPE_IS_STRING_CONCAT = False
 
@@ -129,6 +133,7 @@ class MySQL(Dialect):
             "LONGTEXT": TokenType.LONGTEXT,
             "MEDIUMBLOB": TokenType.MEDIUMBLOB,
             "MEDIUMTEXT": TokenType.MEDIUMTEXT,
+            "MEDIUMINT": TokenType.MEDIUMINT,
             "MEMBER OF": TokenType.MEMBER_OF,
             "SEPARATOR": TokenType.SEPARATOR,
             "START": TokenType.BEGIN,
@@ -136,6 +141,7 @@ class MySQL(Dialect):
             "SIGNED INTEGER": TokenType.BIGINT,
             "UNSIGNED": TokenType.UBIGINT,
             "UNSIGNED INTEGER": TokenType.UBIGINT,
+            "YEAR": TokenType.YEAR,
             "_ARMSCII8": TokenType.INTRODUCER,
             "_ASCII": TokenType.INTRODUCER,
             "_BIG5": TokenType.INTRODUCER,
@@ -185,6 +191,8 @@ class MySQL(Dialect):
         COMMANDS = tokens.Tokenizer.COMMANDS - {TokenType.SHOW}
 
     class Parser(parser.Parser):
+        SUPPORTS_USER_DEFINED_TYPES = False
+
         FUNC_TOKENS = {
             *parser.Parser.FUNC_TOKENS,
             TokenType.DATABASE,
@@ -492,6 +500,17 @@ class MySQL(Dialect):
 
             return self.expression(exp.SetItem, this=charset, collate=collate, kind="NAMES")
 
+        def _parse_type(self) -> t.Optional[exp.Expression]:
+            # mysql binary is special and can work anywhere, even in order by operations
+            # it operates like a no paren func
+            if self._match(TokenType.BINARY, advance=False):
+                data_type = self._parse_types(check_func=True, allow_identifiers=False)
+
+                if isinstance(data_type, exp.DataType):
+                    return self.expression(exp.Cast, this=self._parse_column(), to=data_type)
+
+            return super()._parse_type()
+
     class Generator(generator.Generator):
         LOCKING_READS_SUPPORTED = True
         NULL_ORDERING_SUPPORTED = False
@@ -500,6 +519,7 @@ class MySQL(Dialect):
         DUPLICATE_KEY_UPDATE_WITH_SET = False
         QUERY_HINT_SEP = " "
         VALUES_AS_TABLE = False
+        NVL2_SUPPORTED = False
 
         TRANSFORMS = {
             **generator.Generator.TRANSFORMS,
@@ -515,6 +535,7 @@ class MySQL(Dialect):
             exp.GroupConcat: lambda self, e: f"""GROUP_CONCAT({self.sql(e, "this")} SEPARATOR {self.sql(e, "separator") or "','"})""",
             exp.ILike: no_ilike_sql,
             exp.JSONExtractScalar: arrow_json_extract_scalar_sql,
+            exp.JSONKeyValue: json_keyvalue_comma_sql,
             exp.Max: max_or_greatest,
             exp.Min: min_or_least,
             exp.NullSafeEQ: lambda self, e: self.binary(e, "<=>"),
@@ -524,6 +545,7 @@ class MySQL(Dialect):
             exp.StrPosition: strposition_to_locate_sql,
             exp.StrToDate: _str_to_date_sql,
             exp.StrToTime: _str_to_date_sql,
+            exp.Stuff: rename_func("INSERT"),
             exp.TableSample: no_tablesample_sql,
             exp.TimeStrToUnix: rename_func("UNIX_TIMESTAMP"),
             exp.TimeStrToTime: lambda self, e: self.sql(exp.cast(e.this, "datetime", copy=True)),