improper isolation (overuse of system properties) allows Lucene apps to clobber
each other
--
Key: LUCENE-670
URL: http://issues.apache.org/jira/browse/LUCENE-670
"What's the point of using a sorted interval list for a category?"
Just terminology first to avoid misunderstanding :), category is "category
field" that can take N valus
Now, the case I am facing goes as follows:
I have category field in 50Mio collection which has more or less uniform
distri
Robert Engels,
I implemented the reopen code you posted, works well, thanks. One thing I am
curious about, are you able to reuse the FieldCache? From what I am seeing, it
is being rebuilt after a commit, which makes the next query slow. Any ideas on
this?
Thanks,
Jason
On Thursday 07 September 2006 10:03, eks dev wrote:
...
>
> on the other note,
> the key for really efficiant matchers will be good SmartMatcherFactory that
picks the best representation for given density/"sortednes".
> The cases I've been able to identify so far:
>
> Very Low density - IntList
As suspected in my first email on this topic, test is wrong, garbage collection
was kicking in at wrong times so the results were convoluted. Ah, old good
predictable c times are long gone.
Anyhow, appologies for this noise (particulary to Yonik). The real speed
relationships are attached now,
The same effect showed-up on athlon CPU... The only diference between your
PerfTest and what I have been doing is that your loop is as tight as it goes.
*Maybe* worth of investigating a bit futher as my tests look a bit closer to
the real usage scenarios.
I will try to identify cases where Ope
finalize()-methods of FSDirectory.FSIndexInput and FSDirectory.FSIndexOutput
try to close already closed file
-
Key: LUCENE-669
URL: http://issues.apache.o
