You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Aleksander Eskilson (JIRA)" <ji...@apache.org> on 2018/09/21 14:11:00 UTC
[jira] [Comment Edited] (SPARK-22739) Additional Expression Support
for Objects
[ https://issues.apache.org/jira/browse/SPARK-22739?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16623665#comment-16623665 ]
Aleksander Eskilson edited comment on SPARK-22739 at 9/21/18 2:10 PM:
----------------------------------------------------------------------
[~cloud_fan], could you perhaps link here the Spark issue (if any) and PR that committed built-in Avro support to Spark?
I would like to take a look at how Avro support for Datasets that was created through the PR on this ticket, and an additional PR in Spark-Avro (see [#217|https://github.com/databricks/spark-avro/pull/217]), might be folded in to the new code. I would imagine that process would be including the same expressions this ticket would have included, and then the new AvroEncoder that was to be included in the Spark-Avro project.
Happy to see hear Avro will be included in Spark-proper now!
cc: [~marmbrus]
was (Author: aeskilson):
[~cloud_fan], could you perhaps link here the Spark issue (if any) and PR that committed built-in Avro support to Spark?
I would like to take a look at how Avro support for Datasets that was created through the PR on this ticket, and an additional PR in Spark-Avro (see [#217|https://github.com/databricks/spark-avro/pull/217]), might be folded in to the new code. I would imagine that process would be including the same expressions this ticket would have included, and then the new AvroEncoder that was to be included in the Spark-Avro project.
cc: [~marmbrus]
> Additional Expression Support for Objects
> -----------------------------------------
>
> Key: SPARK-22739
> URL: https://issues.apache.org/jira/browse/SPARK-22739
> Project: Spark
> Issue Type: New Feature
> Components: SQL
> Affects Versions: 2.3.0
> Reporter: Aleksander Eskilson
> Priority: Major
>
> Some discussion in Spark-Avro [1] motivates additions and minor changes to the {{Objects}} Expressions API [2]. The proposed changes include
> * a generalized form of {{initializeJavaBean}} taking a sequence of initialization expressions that can be applied to instances of varying objects
> * an object cast that performs a simple Java type cast against a value
> * making {{ExternalMapToCatalyst}} public, for use in outside libraries
> These changes would facilitate the writing of custom encoders for varying objects that cannot already be readily converted to a statically typed dataset by a JavaBean encoder (e.g. Avro).
> [1] -- https://github.com/databricks/spark-avro/pull/217#issuecomment-342599110
> [2] --
> https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/objects/objects.scala
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org