[
https://issues.apache.org/jira/browse/SPARK-5327?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14283436#comment-14283436
]
Cheng Lian commented on SPARK-5327:
-----------------------------------
Would like to add that the reason why I'd like to run tests under 0.12.0, is
that code paths in the Hive shim layer differ for 0.12.0 and 0.13.1, especially
for Hive UDF and Thrift server related parts.
> HiveCompatibilitySuite fails when executed against Hive 0.12.0
> --------------------------------------------------------------
>
> Key: SPARK-5327
> URL: https://issues.apache.org/jira/browse/SPARK-5327
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 1.3.0, 1.2.1
> Reporter: Cheng Lian
> Assignee: Adrian Wang
>
> Git commit: e7884bc950a374408959b6118efe2c62fbe50608
> Run the following SBT session to reproduce:
> {code}
> $ ./build/sbt -Pyarn,hive,hive-thriftserver,hive-0.12.0,hadoop-2.4,scala-2.10
> -Dhadoop.version=2.4.1
> ...
> > hive/test-only *.HiveCompatibilitySuite -- -z create_view_translate
> ...
> [info] - create_view_translate *** FAILED *** (9 seconds, 216 milliseconds)
> [info] Failed to execute query using catalyst:
> [info] Error: Failed to parse: SELECT `items`.`id`,
> items`items`.`info`info['price'] FROM `default`.`items`
> [info] org.apache.spark.sql.hive.HiveQl$ParseException: Failed to parse:
> SELECT `items`.`id`, items`items`.`info`info['price'] FROM `default`.`items`
> [info] at
> org.apache.spark.sql.hive.HiveQl$.createPlan(HiveQl.scala:249)
> [info] at
> org.apache.spark.sql.hive.HiveQl$.createPlanForView(HiveQl.scala:275)
> [info] at
> org.apache.spark.sql.hive.HiveMetastoreCatalog.lookupRelation(HiveMetastoreCatalog.scala:151)
> ...
> {code}
> Seems that something went wrong when dealing with nested fields. Hive 0.13.1
> is OK.
> There are some other test cases also fail when executed against Hive 0.12.0.
> Will list them later.
> *UPDATE*
> Test cases that should be blacklisted for Hive 0.12.0 (because they use new
> features introduced in 0.13.1)
> - {{udf_to_double}}
> - {{udf_pmod}}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
