Details
-
New Feature
-
Status: Closed
-
Major
-
Resolution: Fixed
-
0.7.0
-
None
Description
"date_sub" function does not accept string as input in Drill, however it does in Hive.
This different behavior of the function will make customer re-write their query to use "cast as date".
Minimum reproduce :
0: jdbc:drill:zk=local> select date_sub('2014-11-12 16:45:22',15) from dfs.tmp.`drilltest/test.csv` ; Query failed: Failure while running fragment., Invalid format: "2014-11-12 16:45:22" is malformed at "14-11-12 16:45:22" [ 9a6f18da-eb1e-4d91-879a-8d9d528efd59 on 10.250.0.115:31010 ] (java.lang.IllegalArgumentException) Invalid format: "2014-11-12 16:45:22" is malformed at "14-11-12 16:45:22" org.joda.time.format.DateTimeFormatter.parseDateTime():873 org.apache.drill.exec.test.generated.ProjectorGen23.doSetup():63 org.apache.drill.exec.test.generated.ProjectorGen23.setup():97 org.apache.drill.exec.physical.impl.project.ProjectRecordBatch.setupNewSchema():427 org.apache.drill.exec.physical.impl.project.ProjectRecordBatch.buildSchema():270 org.apache.drill.exec.physical.impl.validate.IteratorValidatorBatchIterator.buildSchema():80 org.apache.drill.exec.physical.impl.ScreenCreator$ScreenRoot.buildSchema():95 org.apache.drill.exec.work.fragment.FragmentExecutor.run():111 org.apache.drill.exec.work.WorkManager$RunnableWrapper.run():249 .......():0 Error: exception while executing query: Failure while executing query. (state=,code=0)
Comparing to Hive which is good.
0: jdbc:hive2://n1a:10000/default> select date_sub('2014-11-12 16:45:22',15) from passwords limit 1 ;
+-------------+
| _c0 |
+-------------+
| 2014-10-28 |
+-------------+
1 row selected (6.568 seconds)
Workaround in Drill:
0: jdbc:drill:zk=local> select date_sub(cast('2014-11-12 16:45:22' as date),15) from dfs.tmp.`drilltest/test.csv` ; +------------+ | EXPR$0 | +------------+ | 2014-10-28 | +------------+ 1 row selected (0.082 seconds)