[sql lab] extract Hive error messages (#5495)

* [sql lab] extract Hive error messages

So pyhive returns an exception object with a stringified thrift error
object. This PR uses a regex to extract the errorMessage portion of that
string.

* Unit test

(cherry picked from commit 41286b7)

superset/db_engine_specs.py

@@ -1071,10 +1071,10 @@ def adjust_database_uri(cls, uri, selected_schema=None):
 
     @classmethod
     def extract_error_message(cls, e):
-        try:
-            msg = e.message.status.errorMessage
-        except Exception:
-            msg = str(e)
+        msg = str(e)
+        match = re.search('errorMessage="(.*)", ', msg)
+        if match:
+            msg = match.group(1)
         return msg
 
     @classmethod

tests/db_engine_specs_test.py

@@ -5,6 +5,7 @@
 from __future__ import unicode_literals
 
 import textwrap
+from six import text_type
 
 from superset.db_engine_specs import (
     BaseEngineSpec, HiveEngineSpec, MssqlEngineSpec,
@@ -86,6 +87,23 @@ def test_job_2_launched_stage_2_stages_progress(self):
         """.split('\n')  # noqa ignore: E501
         self.assertEquals(60, HiveEngineSpec.progress(log))
 
+    def test_hive_error_msg(self):
+        msg = (
+            '{...} errorMessage="Error while compiling statement: FAILED: '
+            'SemanticException [Error 10001]: Line 4'
+            ':5 Table not found \'fact_ridesfdslakj\'", statusCode=3, '
+            'sqlState=\'42S02\', errorCode=10001)){...}')
+        self.assertEquals((
+            'Error while compiling statement: FAILED: '
+            'SemanticException [Error 10001]: Line 4:5 '
+            "Table not found 'fact_ridesfdslakj'"),
+            HiveEngineSpec.extract_error_message(Exception(msg)))
+
+        e = Exception("Some string that doesn't match the regex")
+        self.assertEquals(
+            text_type(e),
+            HiveEngineSpec.extract_error_message(e))
+
     def get_generic_database(self):
         return Database(sqlalchemy_uri='mysql://localhost')