vfq-cdp-operators 0.0.1__tar.gz

vfq_cdp_operators-0.0.1/PKG-INFO

@@ -0,0 +1,8 @@
+Metadata-Version: 2.4
+Name: vfq_cdp_operators
+Version: 0.0.1
+Summary: Custom Hive Operator for CDE
+Author-email: Data Team <admin@example.com>
+Requires-Python: >=3.7
+Requires-Dist: apache-airflow-providers-common-sql==1.7.0
+Requires-Dist: apache-airflow>=2.0

vfq_cdp_operators-0.0.1/pyproject.toml

@@ -0,0 +1,20 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "vfq_cdp_operators"
+version = "0.0.1"
+authors = [
+  { name="Data Team", email="admin@example.com" },
+]
+description = "Custom Hive Operator for CDE"
+readme = "README.md"
+requires-python = ">=3.7"
+dependencies = [
+    "apache-airflow>=2.0",
+    "apache-airflow-providers-common-sql==1.7.0"
+]
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/vfq_operators"]

vfq_cdp_operators-0.0.1/src/vfq_operators/__init__.py

@@ -0,0 +1 @@
+from .hive_custom import HiveSQLExecuteQueryOperator

vfq_cdp_operators-0.0.1/src/vfq_operators/hive_custom.py

@@ -0,0 +1,25 @@
+from airflow.providers.common.sql.operators.sql import SQLExecuteQueryOperator
+
+class HiveSQLExecuteQueryOperator(SQLExecuteQueryOperator):
+    """
+    Custom Hive Operator that automatically strips trailing semicolons 
+    from split statements to avoid Hive ParseExceptions.
+    """
+    def execute(self, context):
+        # 1. If SQL is passed as a raw string (file content or raw SQL), split manually
+        if isinstance(self.sql, str):
+            # Split by semicolon, clean whitespace, and remove the trailing semicolon
+            self.sql = [
+                s.strip().rstrip(';') 
+                for s in self.sql.split(';') 
+                if s.strip()
+            ]
+            # We handled splitting manually, so turn off the default splitter
+            self.split_statements = False
+            
+        # 2. If SQL is already a list (rare, but possible), just clean it
+        elif isinstance(self.sql, list):
+            self.sql = [s.strip().rstrip(';') for s in self.sql if s.strip()]
+
+        # 3. Pass the cleaned SQL to the standard operator logic
+        return super().execute(context)