lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Miller <>
Subject Re: [jira] Commented: (LUCENE-1243) A few new benchmark tasks
Date Tue, 02 Sep 2008 19:00:29 GMT
And all that has been added is a better way to do the sort testing - its 
a new doc maker that you lets pick a range of random numbers to generate 
in a known sort field. The range of random ints to be generated can be 
specified. Its fairly simple, but its a start that works, and if we need 
more than the basics that are provided, they can be added easily later. 
Unless there are suggestions for any changes wanted, I think its all 
good to go in.

Grant Ingersoll (JIRA) wrote:
>     [
> Grant Ingersoll commented on LUCENE-1243:
> -----------------------------------------
> My last look at it seemed like it was in good shape, but that was a  
> while ago.
>> A few new benchmark tasks
>> -------------------------
>>                 Key: LUCENE-1243
>>                 URL:
>>             Project: Lucene - Java
>>          Issue Type: Improvement
>>          Components: contrib/benchmark
>>            Reporter: Mark Miller
>>            Assignee: Grant Ingersoll
>>            Priority: Minor
>>             Fix For: 2.4
>>         Attachments: benchmark-tasks.diff, LUCENE-1243.03.30.2008.patch, LUCENE-1243.patch
>> Some tasks that would be helpful to see added. Might want some expansion, but here
are some basic ones I have been using:
>> CommitIndexTask
>> ReopenReaderTask
>> SearchWithSortTask
>> I do the sort in a similar way that the highlighting was done, but another method
may be better. Just would be great to have sorting.
>> Also, since there is no great field for sorting (reuters date always appears to be
the same) I changed the id field from doc+id to just id. Again maybe not the best solution,
but here I am to get the ball rolling :)

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message