[
https://issues.apache.org/jira/browse/HBASE-8063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13599661#comment-13599661
]
Liang Xie commented on HBASE-8063:
----------------------------------
O, seems there's a trouble here, my two thoughts
# generate keyvalue instances per scan.start/stopRow(seems current KeyValue
class doesn't have any methods like createKeyValueFromRow,right?), then compare
with getFirstKey/getLastKey, maybe we need to handle some details,e.g. kv's
timestamp, type? it's error-prone for me at least
# shortcut MetaKeyComparator scenario, just return true if it's a
MetaKeyComparator instance. since our original idea is to optimize normal user
hfile, not internal system hfile, it should be ok. I prefer to this choice.
[~zjushch], how about you?
> Filter HFiles based on first/last key
> -------------------------------------
>
> Key: HBASE-8063
> URL: https://issues.apache.org/jira/browse/HBASE-8063
> Project: HBase
> Issue Type: Improvement
> Components: Scanners
> Affects Versions: 0.98.0
> Reporter: Liang Xie
> Assignee: Liang Xie
> Attachments: HBASE-8063.txt
>
>
> # hbase-5010 introduced a filter mechanism based on timerange hint
> # this jira is a placeholder just for keyrange hint: currently, low
> layer(hfile reader) has first/last key getter methods already. we should
> utilize this information to pre-filte hfiles while choosing scanners from
> storefiles.
> It's just a raw thought right now, no patch available yet. I'll summit a
> patch asap if it's reasonable, welcome to give some suggestions/comments:)
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
