Feat: add support for IS_ASCII (#4557)

* IS_ASCII support

* linter fix

* mysql is_ascii simplified

Author: pruzko

sqlglot/dialects/mysql.py

@@ -1261,3 +1261,6 @@ def converttimezone_sql(self, expression: exp.ConvertTimezone) -> str:
         def attimezone_sql(self, expression: exp.AtTimeZone) -> str:
             self.unsupported("AT TIME ZONE is not supported by MySQL")
             return self.sql(expression.this)
+
+        def isascii_sql(self, expression: exp.IsAscii) -> str:
+            return f"REGEXP_LIKE({self.sql(expression.this)}, '^[[:ascii:]]*$')"

sqlglot/dialects/oracle.py

@@ -398,3 +398,6 @@ def hint_sql(self, expression: exp.Hint) -> str:
                     expressions.append(self.sql(expression))
 
             return f" /*+ {self.expressions(sqls=expressions, sep=self.QUERY_HINT_SEP).strip()} */"
+
+        def isascii_sql(self, expression: exp.IsAscii) -> str:
+            return f"NVL(REGEXP_LIKE({self.sql(expression.this)}, '^[' || CHR(1) || '-' || CHR(127) || ']*$'), TRUE)"

sqlglot/dialects/postgres.py

@@ -716,3 +716,6 @@ def array_sql(self, expression: exp.Array) -> str:
 
         def computedcolumnconstraint_sql(self, expression: exp.ComputedColumnConstraint) -> str:
             return f"GENERATED ALWAYS AS ({self.sql(expression, 'this')}) STORED"
+
+        def isascii_sql(self, expression: exp.IsAscii) -> str:
+            return f"({self.sql(expression.this)} ~ '^[[:ascii:]]*$')"

sqlglot/dialects/sqlite.py

@@ -300,3 +300,6 @@ def transaction_sql(self, expression: exp.Transaction) -> str:
             this = expression.this
             this = f" {this}" if this else ""
             return f"BEGIN{this} TRANSACTION"
+
+        def isascii_sql(self, expression: exp.IsAscii) -> str:
+            return f"(NOT {self.sql(expression.this)} GLOB CAST(x'2a5b5e012d7f5d2a' AS TEXT))"

sqlglot/dialects/tsql.py

@@ -1267,3 +1267,6 @@ def dpipe_sql(self, expression: exp.DPipe) -> str:
             return self.sql(
                 reduce(lambda x, y: exp.Add(this=x, expression=y), expression.flatten())
             )
+
+        def isascii_sql(self, expression: exp.IsAscii) -> str:
+            return f"(PATINDEX('%[^' + CHAR(0x00) + '-' + CHAR(0x7f) + ']%' COLLATE Latin1_General_BIN, {self.sql(expression.this)}) = 0)"

sqlglot/expressions.py

@@ -5891,6 +5891,10 @@ class Initcap(Func):
     arg_types = {"this": True, "expression": False}
 
 
+class IsAscii(Func):
+    pass
+
+
 class IsNan(Func):
     _sql_names = ["IS_NAN", "ISNAN"]
 

tests/dialects/test_dialect.py

@@ -520,6 +520,19 @@ def test_if_null(self):
             },
         )
 
+    def test_is_ascii(self):
+        self.validate_all(
+            "SELECT IS_ASCII(x)",
+            write={
+                "": "SELECT IS_ASCII(x)",
+                "sqlite": "SELECT (NOT x GLOB CAST(x'2a5b5e012d7f5d2a' AS TEXT))",
+                "mysql": "SELECT REGEXP_LIKE(x, '^[[:ascii:]]*$')",
+                "postgres": "SELECT (x ~ '^[[:ascii:]]*$')",
+                "tsql": "SELECT (PATINDEX('%[^' + CHAR(0x00) + '-' + CHAR(0x7f) + ']%' COLLATE Latin1_General_BIN, x) = 0)",
+                "oracle": "SELECT NVL(REGEXP_LIKE(x, '^[' || CHR(1) || '-' || CHR(127) || ']*$'), TRUE)",
+            },
+        )
+
     def test_nvl2(self):
         self.validate_all(
             "SELECT NVL2(a, b, c)",