[
https://issues.apache.org/jira/browse/SPARK-21365?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16080665#comment-16080665
]
Apache Spark commented on SPARK-21365:
--------------------------------------
User 'HyukjinKwon' has created a pull request for this issue:
https://github.com/apache/spark/pull/18590
> Deduplicate logics parsing DDL-like type definition
> ---------------------------------------------------
>
> Key: SPARK-21365
> URL: https://issues.apache.org/jira/browse/SPARK-21365
> Project: Spark
> Issue Type: Improvement
> Components: PySpark
> Affects Versions: 2.2.0
> Reporter: Hyukjin Kwon
>
> It looks we duplicate
> https://github.com/apache/spark/blob/d492cc5a21cd67b3999b85d97f5c41c3734b1ba3/python/pyspark/sql/types.py#L823-L845
> logic for parsing DDL-like type definitions.
> There are also two more points here:
> - This does not support "field type" but "field: type".
> - This does not support nested schemas. For example as below:
> {code}
> >>> spark.createDataFrame([[[1]]], "struct<a: struct<b: int>>").show()
> ...
> ValueError: The strcut field string format is: 'field_name:field_type', but
> got: a: struct<b: int>
> {code}
> {code}
> >>> spark.createDataFrame([[[1]]], "a: struct<b: int>").show()
> ...
> ValueError: The strcut field string format is: 'field_name:field_type', but
> got: a: struct<b: int>
> {code}
> {code}
> >>> spark.createDataFrame([[[1]]], "a int").show()
> ...
> ValueError: Could not parse datatype: a int
> {code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
