[
https://issues.apache.org/jira/browse/HBASE-17375?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15779832#comment-15779832
]
ramkrishna.s.vasudevan commented on HBASE-17375:
------------------------------------------------
So what is your feedback on Prefix tree other than stableness. Are you doing
some POC now with it? What was your comparison with other encoding algos? You
are getting better perf with random access? I think your feedback is important
so that if really users find it interesting, then as a community, it is bound
to us to make it stable so that users could use it efficiently.
Thanks [~baibaichen].
I could help here but I just forgot what is there inside prefix. So may need
some time to get back to it.
> PrefixTreeArrayReversibleScanner#previousRowInternal doesn't work correctly
> ---------------------------------------------------------------------------
>
> Key: HBASE-17375
> URL: https://issues.apache.org/jira/browse/HBASE-17375
> Project: HBase
> Issue Type: Bug
> Affects Versions: 1.1.7, 0.98.23, 0.98.24
> Reporter: Chang chen
> Attachments: row trie example.PNG
>
>
> Recently, we find our hbase compaction thread never end. Assume we have
> following cells:
> {quote}
> <A,a> 1
> <A,v> 1
> <Aaeeee,a> 1
> <Aaeeee,v> 1
> <Abc,a> 1
> <Abc,v> 1
> <Abde,a> 1
> <Abde,v> 1
> {quote}
> If we encode above datas into prefix tree block, then it looks like:
> !row trie example.PNG!
> Assume the current row is {color:red}Abc{color} (e.g. the current row node is
> 4), then the previous row should be *Aaeeee* (e.g. 2). However
> previousRowInternal return {color:red}A{color}(e.g. 1)
> After investigation, I believe it's the bug of
> PrefixTreeArrayReversibleScanner#previousRowInternal.
> {code}
> private boolean previousRowInternal() {
> //...
> while (!beforeFirst) {
> //....
> // what if currentRowNode is nub?
> if (currentRowNode.hasOccurrences()) {// escape clause
> currentRowNode.resetFanIndex();
> return true;// found some values
> }
> }
> {code}
> currentRowNode.hasOccurrences() only test whether it has cell or not. But in
> the case of currentRowNode.isNub() is true, previousRowInternal should
> follow the previous fan instead of return.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
