[
https://issues.apache.org/jira/browse/SPARK-14918?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15258191#comment-15258191
]
Cheng Lian commented on SPARK-14918:
------------------------------------
Hit this issue while reviewing https://github.com/apache/spark/pull/12222
However, checked all usages of {{TablePartitionSpec}} throughout current master
branch, none of them assumes that {{TablePartitionSpec}} preserves the order. I
tend to fix the problematic PR by stop relying on the wrong assumption.
> ExternalCatalog.TablePartitionSpec doesn't preserve partition column order
> --------------------------------------------------------------------------
>
> Key: SPARK-14918
> URL: https://issues.apache.org/jira/browse/SPARK-14918
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 2.0.0
> Reporter: Cheng Lian
> Assignee: Cheng Lian
>
> Equivalent entity of {{ExternalCatalog.TablePartitionSpec}} in Hive is a
> {{LinkedHashMap}} returned by {{Partition.getSpec()}}, which preserves
> partition column order.
> However, we are using a {{scala.immutable.Map}} to store the result, which no
> longer preserves the original order. What makes it worse, Scala specializes
> immutable maps with less than 5 elements. And these specialized versions do
> preserve order, thus hides this issue in test cases since we never use more
> than 4 partition columns.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
