From 41286b7545f8fc6a4a8c9782b169afd62a73e9c1 Mon Sep 17 00:00:00 2001
From: Maxime Beauchemin <maximebeauchemin@gmail.com>
Date: Thu, 26 Jul 2018 15:17:55 -0700
Subject: [PATCH] [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
---
 superset/db_engine_specs.py   |  8 ++++----
 tests/db_engine_specs_test.py | 19 +++++++++++++++++++
 2 files changed, 23 insertions(+), 4 deletions(-)

diff --git a/superset/db_engine_specs.py b/superset/db_engine_specs.py
index cc1345e371a88..5e5a30414ad60 100644
--- a/superset/db_engine_specs.py
+++ b/superset/db_engine_specs.py
@@ -1051,10 +1051,10 @@ 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
diff --git a/tests/db_engine_specs_test.py b/tests/db_engine_specs_test.py
index c85e23a26c023..287114ccd7a0f 100644
--- a/tests/db_engine_specs_test.py
+++ b/tests/db_engine_specs_test.py
@@ -4,6 +4,8 @@
 from __future__ import print_function
 from __future__ import unicode_literals
 
+from six import text_type
+
 from superset.db_engine_specs import (
     BaseEngineSpec, HiveEngineSpec, MssqlEngineSpec,
     MySQLEngineSpec, PrestoEngineSpec,
@@ -84,6 +86,23 @@ 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')