[
https://issues.apache.org/jira/browse/CHUKWA-594?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13059416#comment-13059416
]
Preetam Patil commented on CHUKWA-594:
--------------------------------------
There is a related bug in the HICC java code (ChukwaHbaseStore.java). The
RowFilter in getSeries does a substring matching which means that a filter for
"node1" returns all entries for "node1*".
Here's the original code segment:
{{if(filterByRowKey) {
RowFilter rf = new RowFilter(CompareOp.EQUAL, new
SubstringComparator(rkey));
scan.setFilter(rf);
}
}}
A solution is to use a RegexStringComparator, as in the following (modified)
segment:
{{if(filterByRowKey) {
RowFilter rf = new RowFilter(CompareOp.EQUAL, new
RegexStringComparator("[0-9]+-"+rkey+"$"));
scan.setFilter(rf);
}
}}
I've tested it to be working okay. If it looks okay, I'll submit a patch soon.
-preetam
> Substring matching in Row (system name) in HICC web interface (function
> getRows of graph_explorer.jsp)
> ------------------------------------------------------------------------------------------------------
>
> Key: CHUKWA-594
> URL: https://issues.apache.org/jira/browse/CHUKWA-594
> Project: Chukwa
> Issue Type: Bug
> Components: User Interface
> Affects Versions: 0.5.0
> Environment: Ubuntu 10.04 amd64 with Java 1.6.0_24
> Reporter: Preetam Patil
> Assignee: Eric Yang
> Priority: Minor
> Labels: javascript
> Attachments: CHUKWA-594.patch
>
>
> In the web interface file graph_explorer.jsp, when querying the rows (system)
> with unique names (function getRows), the row (node) names that are
> substrings of existing names are dropped from the list provided by the
> webserver. For example, if 'node1' and 'node11' are part of the list, 'node1'
> is dropped and not displayed.
> I'm not a Javascript expert--perhaps the call to ": contains" matches
> substrings?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
