chore: improve DML check

apache · Sep 27, 2024 · 4b74ebf · 4b74ebf
1 parent 4f5f8fd
commit 4b74ebf
Show file tree

Hide file tree

Showing 4 changed files with 28 additions and 11 deletions.
diff --git a/pyproject.toml b/pyproject.toml
@@ -89,7 +89,7 @@ dependencies = [
     "slack_sdk>=3.19.0, <4",
     "sqlalchemy>=1.4, <2",
     "sqlalchemy-utils>=0.38.3, <0.39",
-    "sqlglot>=23.0.2,<24",
+    "sqlglot>=25.24.0,<26",
     "sqlparse>=0.5.0",
     "tabulate>=0.8.9, <0.9",
     "typing-extensions>=4, <5",

diff --git a/requirements/base.txt b/requirements/base.txt
@@ -350,7 +350,7 @@ sqlalchemy-utils==0.38.3
     # via
     #   apache-superset
     #   flask-appbuilder
-sqlglot==23.6.3
+sqlglot==25.24.0
     # via apache-superset
 sqlparse==0.5.0
     # via apache-superset

diff --git a/superset/sql_lab.py b/superset/sql_lab.py
@@ -46,6 +46,7 @@
     OAuth2RedirectError,
     SupersetErrorException,
     SupersetErrorsException,
+    SupersetParseError,
 )
 from superset.extensions import celery_app, event_logger
 from superset.models.core import Database
@@ -236,15 +237,27 @@ def execute_sql_statement(  # pylint: disable=too-many-statements, too-many-loca
     # We are testing to see if more rows exist than the limit.
     increased_limit = None if query.limit is None else query.limit + 1
 
-    parsed_statement = SQLStatement(sql_statement, engine=db_engine_spec.engine)
-    if parsed_statement.is_mutating() and not database.allow_dml:
-        raise SupersetErrorException(
-            SupersetError(
-                message=__("Only SELECT statements are allowed against this database."),
-                error_type=SupersetErrorType.DML_NOT_ALLOWED_ERROR,
-                level=ErrorLevel.ERROR,
+    if not database.allow_dml:
+        try:
+            parsed_statement = SQLStatement(sql_statement, engine=db_engine_spec.engine)
+            disallowed = parsed_statement.is_mutating()
+        except SupersetParseError:
+            # if we fail to parse teh query, disallow by default
+            disallowed = True
+
+        if disallowed:
+            raise SupersetErrorException(
+                SupersetError(
+                    message=__(
+                        "This database does not allow for DDL/DML, and the query "
+                        "could not be parsed to confirm it is a read-only query. Please "
+                        "contact your administrator for more assistance."
+                    ),
+                    error_type=SupersetErrorType.DML_NOT_ALLOWED_ERROR,
+                    level=ErrorLevel.ERROR,
+                )
             )
-        )
+
     if apply_ctas:
         if not query.tmp_table_name:
             start_dttm = datetime.fromtimestamp(query.start_time)

diff --git a/tests/integration_tests/sqllab_tests.py b/tests/integration_tests/sqllab_tests.py
@@ -137,7 +137,11 @@ def test_sql_json_dml_disallowed(self):
         assert data == {
             "errors": [
                 {
-                    "message": "Only SELECT statements are allowed against this database.",
+                    "message": (
+                        "This database does not allow for DDL/DML, and the query "
+                        "could not be parsed to confirm it is a read-only query. Please "
+                        "contact your administrator for more assistance."
+                    ),
                     "error_type": SupersetErrorType.DML_NOT_ALLOWED_ERROR,
                     "level": ErrorLevel.ERROR,
                     "extra": {