[
https://issues.apache.org/jira/browse/DRILL-5838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16203679#comment-16203679
]
ASF GitHub Bot commented on DRILL-5838:
---------------------------------------
Github user vvysotskyi commented on the issue:
https://github.com/apache/drill/pull/972
@paul-rogers, Thanks for the so informative comments, I completely agree
with the convention that you have proposed.
The problem which this PR is addressed, partially connected to the third
point:
> When calling back out to the external system, the interface (plugin) must
convert from Drill format back out to the external system format.
MaprDB uses `FieldPath` class to represent the field. Since the string,
that is received in the method `JsonConditionBuilder.createJsonScanSpec()` from
`SchemaPath` instance will be parsed using `org.ojai.FieldPath.parseFrom()`
method, to avoid any ambiguousness, I converted `SchemaPath` to `FieldPath` and
then called `FieldPath.asPathString()`.
> Fix MaprDB filter pushdown for the case of nested field (reg. of DRILL-4264)
> ----------------------------------------------------------------------------
>
> Key: DRILL-5838
> URL: https://issues.apache.org/jira/browse/DRILL-5838
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - MapRDB
> Affects Versions: 1.11.0
> Reporter: Volodymyr Vysotskyi
> Assignee: Volodymyr Vysotskyi
>
> After the changes in DRILL-4264, filter for MaprDB tables operate only with
> root field name, but {{MaprDBJsonRecordReader}} expects to use condition with
> full field path. In MaprDB classes compares the results of calling
> {{org.ojai.FieldPath.asPathString()}} method for the fields from project list
> and from filter list.
> So current code works only for cases when compares root fields, but not
> nested.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)