[
https://issues.apache.org/jira/browse/HIVE-19137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16442812#comment-16442812
]
Eugene Koifman commented on HIVE-19137:
---------------------------------------
Go ahead and take it if you want. The issue is not during writing - if you use
API to read the footer of the orc file, the info is there. The issue is that
orcfiledump only prints text data and I made this field binary for some reason.
You can look at OrcRecordUpdater.KeyIndexBuilder for example and UTF8 encode
version info in the same fashion - this should make it visible in orcfiledump
> orcfiledump doesn't print hive.acid.version value
> -------------------------------------------------
>
> Key: HIVE-19137
> URL: https://issues.apache.org/jira/browse/HIVE-19137
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Affects Versions: 3.0.0
> Reporter: Eugene Koifman
> Assignee: Eugene Koifman
> Priority: Major
>
> HIVE-18659 added hive.acid.version in the file footer.
> orcfiledump prints something like
> {noformat}
> User Metadata:
> hive.acid.key.index=1,536870912,1;
> hive.acid.stats=2,0,0
> hive.acid.version=
> {noformat}
> probably because
> {noformat}
> public static void setAcidVersionInDataFile(Writer writer) {
> //so that we know which version wrote the file
> ByteBuffer bf = ByteBuffer.allocate(4).putInt(ORC_ACID_VERSION);
> bf.rewind(); //don't ask - some ByteBuffer weridness. w/o this, empty
> buffer is written
> writer.addUserMetadata(ACID_VERSION_KEY, bf);
> }
> {noformat}
> use
> {{UTF8.encode())}} instead
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
