[
https://issues.apache.org/jira/browse/SPARK-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14496635#comment-14496635
]
Yin Huai commented on SPARK-4967:
---------------------------------
For new parquet, because we only accept a single path (we also use hadoop
{{Path}} to set the input path, so Hadoop will escape comma), I think we will
not hit it for now.
> File name with comma will cause exception for SQLContext.parquetFile
> --------------------------------------------------------------------
>
> Key: SPARK-4967
> URL: https://issues.apache.org/jira/browse/SPARK-4967
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Reporter: Cheng Hao
> Priority: Minor
>
> {code}
> test("simple projection with comma in file name") {
> val tmpFile = util.getTempFilePath("t,t").getCanonicalFile
> Utils.registerShutdownDeleteDir(tmpFile)
> // Test if "," in the file name can be correctly handled
> sparkContext.parallelize((1, 1) ::
> Nil).saveAsParquetFile(tmpFile.getCanonicalPath)
> assert(parquetFile(tmpFile.getCanonicalPath).first() === Row(1, 1))
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
