[ https://issues.apache.org/jira/browse/CRUNCH-569?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Micah Whitacre updated CRUNCH-569: ---------------------------------- Attachment: CRUNCH-569.patch Ok took a quick pass a handling the deep copy when persisting. It makes the original test pass which is good but I don't think the work is necessarily complete as I'm guessing similar changes for the other impls in org.apache.crunch.impl.spark.collect are still needed. [~joshwills] does this seem to align with what you were thinking enough to try and carry forward or is there a better way? > Don't reuse avros on SparkPipeline cache() > ------------------------------------------ > > Key: CRUNCH-569 > URL: https://issues.apache.org/jira/browse/CRUNCH-569 > Project: Crunch > Issue Type: Bug > Components: Spark > Affects Versions: 0.12.0 > Reporter: Nithin Asokan > Attachments: CRUNCH-569.patch, CRUNCH-569_test.patch > > > A cached PCollection/PTable built by Avros using SparkPipeline seem to reuse > avro objects. Here is a test that shows this behavior > https://gist.github.com/nasokan/531b4ff9bf827d0835ab > I would expect the output of this program to create a pair with same key, > value. However, this produces Pair with different key value. I have tested > this with text file input source and it works as expected. Removing cache() > produces expected result. > Mailing list: > http://mail-archives.apache.org/mod_mbox/crunch-user/201510.mbox/%3CCAH29n6OWD%2BSoa0A88NwaiurZcN%2B2Ow-s_zk3b%2BcKoUMNMO8uuA%40mail.gmail.com%3E -- This message was sent by Atlassian JIRA (v6.3.4#6332)