[
https://issues.apache.org/jira/browse/HAWQ-53?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14968159#comment-14968159
]
ASF GitHub Bot commented on HAWQ-53:
------------------------------------
GitHub user shivzone opened a pull request:
https://github.com/apache/incubator-hawq/pull/33
HAWQ-53. Fix to handling nulls in complex data types in avro via pxf
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/apache/incubator-hawq HAWQ-53
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-hawq/pull/33.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #33
----
commit 098e241db115d8cd9fa3f710f546fb300afe9db0
Author: Shivram Mani <[email protected]>
Date: 2015-10-21T22:44:21Z
HAWQ-53. Fix to handling nulls in complex data types in avro via pxf
----
> Avro Union(complex) type containing nullable type and using null as default
> failed
> ----------------------------------------------------------------------------------
>
> Key: HAWQ-53
> URL: https://issues.apache.org/jira/browse/HAWQ-53
> Project: Apache HAWQ
> Issue Type: Bug
> Components: PXF
> Reporter: Goden Yao
> Assignee: Shivram Mani
> Attachments: meetme.avsc, meetme_example.json
>
>
> *Current Behavior*
> Create pxf external table with an avro schema which contains complex nullable
> type with null as default value would fail.
> *Expected Behavior*
> The complex avro types should be properly mapped to HAWQ types and create
> table should have no error with data loaded.
> *Analysis*
> From code review
> [AvroResolver.java|https://github.com/apache/incubator-hawq/blob/master/pxf/pxf-hdfs/src/main/java/com/pivotal/pxf/plugins/hdfs/AvroResolver.java],
> we didn't specify null situation for complex types:
> - Array
> - Map
> - Record
> when using union type , which contains a nullable type(union) and use null as
> default - I think complex type with nullable type in value schema and with
> null as default is the problem.
> {code:javascript}
> ...
> "name" : "meta__kvpairs",
> "type" : [ "null", {
> "type" : "map",
> "values" : [ "null", "string" ],
> "default" : null
> } ],
> "default" : null
> },
> ...
> {code}
> Create external table from pxf to read it:
> {code:sql}
> "CREATE EXTERNAL TABLE meetmeevent_hdfs(
> sourcetimestamp bigint,
> sourceagent varchar(200),
> sourceip_address varchar(20),
> meetmeuser_id numeric(15,0),
> meetmeinterested_uid numeric(15,0),
> meetmevote varchar(4),
> meetmeis_match numeric(3,0),
> meetmenetwork_score numeric (10,0),
> meetmeresponsiveness_score numeric (10,0),
> meetmesession_id varchar(32),
> meetmefriends numeric(1,0),
> meetmeprevious_view varchar(32),
> meetmemodel numeric(15,0),
> meetmescore numeric(15,0),
> meetmemethod numeric(1,0),
> meetmecontributions varchar(120),
> meetmeclicksource varchar(20),
> meetmeclickaction varchar(20),
> meetmeprofileview_ts bigint,
> meetmeplatform varchar(20),
> metatopic_name varchar(100),
> metarequest_user_agent varchar(200),
> metarequest_session_id varchar(50),
> metarequest_id bigint,
> metakvpairs varchar(1000),
> meta_handlers varchar(1000),
> dt varchar(6),
> hour varchar(2)
> )
> LOCATION
> ('pxf://dahdp2nn01.tag-dev.com:50070/data/ramblas/event_log/meetme/20150824/12/s_meetme.20.0.3492572.11600054147.1440442800000.avro?PROFILE=Avro&DATA-SCHEMA=/data/ramblas/schema/meetme')
> FORMAT 'CUSTOM' (formatter='pxfwritable_import');
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
