[
https://issues.apache.org/jira/browse/SPARK-8056?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14574847#comment-14574847
]
Ilya Ganelin edited comment on SPARK-8056 at 6/5/15 5:18 PM:
-------------------------------------------------------------
[~rxin] Sounds good :). Where would you suggest adding a test for StructType
creation? Not sure where it quite fits in the grand scheme of things.
With regards to also supporting a string for simple types, I think it's safer
to enforce usage of DataType since I think the intent is for the SQL schema to
be strictly typed. Were you suggesting that we allow passing "int" or "long" as
the type argument or for us to infer it automatically by parsing the string?
That approach seems a little more dangerous.
was (Author: ilganeli):
[~rxin] Sounds good :). Where would you suggest adding a test for StructType
creation? Not sure where it quite fits in the grand scheme of things.
With regards to also supporting a string for simple types, I think it's safer
to enforce usage of DataType since the SQL schema should be strictly typed.
Were you suggesting that we allow passing "int" or "long" as the type argument
or for us to infer it automatically by parsing the string? That approach seems
a little more dangerous.
> Design an easier way to construct schema for both Scala and Python
> ------------------------------------------------------------------
>
> Key: SPARK-8056
> URL: https://issues.apache.org/jira/browse/SPARK-8056
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Reporter: Reynold Xin
>
> StructType is fairly hard to construct, especially in Python.
--
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
