[
https://issues.apache.org/jira/browse/MAPREDUCE-5937?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15221335#comment-15221335
]
Akira AJISAKA commented on MAPREDUCE-5937:
------------------------------------------
Anyway, I think we need to avoid returning duplicate Strings in
AbstractCounters#getGroupNames.
{code}
if(isFGroup ? fgroups.containsKey(newGroup) :
groups.containsKey(newGroup)) {
deprecated.add(entry.getKey());
}
{code}
We should avoid adding {{entry.getKey()}} to {{deprecated}} when
{{entry.getKey()}} is already in {{fgroups.keySet()}} or {{groups.keySet()}}.
> hadoop/mapred job -history <history_file> shows the counters twice in the
> output.
> ---------------------------------------------------------------------------------
>
> Key: MAPREDUCE-5937
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5937
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Affects Versions: 2.2.0, 2.7.2
> Reporter: Jinghui Wang
> Assignee: Andres Perez
> Labels: BB2015-05-TBR
> Attachments: MAPREDUCE-5937-branch-2.7-02.patch,
> MAPREDUCE-5937-branch-2.7.2.002.patch, MAPREDUCE-5937.patch,
> job_history_cli_sample.out
>
>
> HiistoryView#printCounters method uses AbstractCounter#getGroupNames, which
> includes legacy groups can cause duplicates on CLI output.
> See attached example output.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
