[
https://issues.apache.org/jira/browse/SPARK-48100?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Shujing Yang updated SPARK-48100:
---------------------------------
Description:
Previously, the XML parser can't skip nested structure data fields effectively
when they were not selected in the schema. For instance, in the below example,
`df.select("struct2").collect()` returns `Seq(null)` as `struct1` wasn't
effectively skipped. This PR fixes this issue.
{code:java}
<ROW>
<struct1>
<innerStruct><field1>1</field1></innerStruct>
</struct1>
<struct2>
<field2>2</field2>
</struct2>
</ROW>{code}
was:
Previously, the XML parser can't skip nested structure data fields when they
were not selected in the schema. For instance, in the below example,
`df.select("struct2").collect()` returns `Seq(null)` as `struct1` wasn't
effectively skipped. This PR fixes this issue.
{code:java}
<ROW>
<struct1>
<innerStruct><field1>1</field1></innerStruct>
</struct1>
<struct2>
<field2>2</field2>
</struct2>
</ROW>{code}
> [SQL][XML] Fix issues in skipping nested structure fields not selected in
> schema
> --------------------------------------------------------------------------------
>
> Key: SPARK-48100
> URL: https://issues.apache.org/jira/browse/SPARK-48100
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 4.0.0
> Reporter: Shujing Yang
> Priority: Major
> Labels: pull-request-available
>
> Previously, the XML parser can't skip nested structure data fields
> effectively when they were not selected in the schema. For instance, in the
> below example, `df.select("struct2").collect()` returns `Seq(null)` as
> `struct1` wasn't effectively skipped. This PR fixes this issue.
> {code:java}
> <ROW>
> <struct1>
> <innerStruct><field1>1</field1></innerStruct>
> </struct1>
> <struct2>
> <field2>2</field2>
> </struct2>
> </ROW>{code}
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
