[
https://issues.apache.org/jira/browse/PARQUET-328?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14617332#comment-14617332
]
Ryan Blue commented on PARQUET-328:
-----------------------------------
File system handles are cached by the Hadoop FS API, so it is safe to call
FileSystem.get every time. I agree with Jacques that we shouldn't cache in
Parquet, though I don't like the idea of relying on a user to pass in the
correct file system. That causes errors when they pass a FS that doesn't match.
> ParquetReader should cache FileSystem object
> --------------------------------------------
>
> Key: PARQUET-328
> URL: https://issues.apache.org/jira/browse/PARQUET-328
> Project: Parquet
> Issue Type: Bug
> Reporter: Tianshuo Deng
>
> ParquetReader should cache FileSystem object instead of re-initing FileSystem
> every time
> https://github.com/apache/parquet-mr/blob/master/parquet-hadoop/src/main/java/org/apache/parquet/hadoop/ParquetFileReader.java#L385
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
