[airflow] branch main updated: Validate connection host field for Sqoop connection (#32968)

[potiuk]

This is an automated email from the ASF dual-hosted git repository.

potiuk pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/airflow.git


The following commit(s) were added to refs/heads/main by this push:
     new ddcd30e7c7 Validate connection host field for Sqoop connection (#32968)
ddcd30e7c7 is described below

commit ddcd30e7c7f5daeab5f74fb3224a4d5e33cec95d
Author: Pankaj Koti <pankajkoti...@gmail.com>
AuthorDate: Mon Jul 31 18:48:42 2023 +0530

    Validate connection host field for Sqoop connection (#32968)
    
    The connection `host` field should not contain a `?` for the
    sqoop connection as it is not intended to include query params
    in the `host` field.
---
 airflow/providers/apache/sqoop/hooks/sqoop.py    |  3 +++
 tests/providers/apache/sqoop/hooks/test_sqoop.py | 15 +++++++++++++++
 2 files changed, 18 insertions(+)

diff --git a/airflow/providers/apache/sqoop/hooks/sqoop.py 
b/airflow/providers/apache/sqoop/hooks/sqoop.py
index b46984c9f0..297c567c19 100644
--- a/airflow/providers/apache/sqoop/hooks/sqoop.py
+++ b/airflow/providers/apache/sqoop/hooks/sqoop.py
@@ -113,6 +113,9 @@ class SqoopHook(BaseHook):
                 raise AirflowException(f"Sqoop command failed: {masked_cmd}")
 
     def _prepare_command(self, export: bool = False) -> list[str]:
+        if "?" in self.conn.host:
+            raise ValueError("The sqoop connection host should not contain a 
'?' character")
+
         sqoop_cmd_type = "export" if export else "import"
         connection_cmd = ["sqoop", sqoop_cmd_type]
 
diff --git a/tests/providers/apache/sqoop/hooks/test_sqoop.py 
b/tests/providers/apache/sqoop/hooks/test_sqoop.py
index 2570b356bf..0075ebec9b 100644
--- a/tests/providers/apache/sqoop/hooks/test_sqoop.py
+++ b/tests/providers/apache/sqoop/hooks/test_sqoop.py
@@ -101,6 +101,16 @@ class TestSqoopHook:
                 extra=None,
             )
         )
+        db.merge_conn(
+            Connection(
+                conn_id="invalid_host_conn",
+                conn_type="mssql",
+                schema="schema",
+                host="rmdbs?query_param1=value1",
+                port=5050,
+                extra=None,
+            )
+        )
 
     @patch("subprocess.Popen")
     def test_popen(self, mock_popen):
@@ -370,3 +380,8 @@ class TestSqoopHook:
         # Case no mssql
         hook = SqoopHook(conn_id="sqoop_test")
         assert f"{hook.conn.host}:{hook.conn.port}/{hook.conn.schema}" in 
hook._prepare_command()
+
+    def test_invalid_host(self):
+        hook = SqoopHook(conn_id="invalid_host_conn")
+        with pytest.raises(ValueError, match="host should not contain a"):
+            hook._prepare_command()