[
https://issues.apache.org/jira/browse/LUCENE-1821?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12746263#action_12746263
]
Tim Smith commented on LUCENE-1821:
-----------------------------------
I started integrating the per-segment searching (removed my hack that was doing
searching on MultiReader)
In order to get my query implementations to work, i had to hold onto my
Searcher in the Weight constructor and add getIndexReaderBase() method to my
IndexSearcher implementation, and this seems to be working well
I had 3 query implementations that were affected:
one used a cache that will be easy to create per segment (will have this use a
per segment cache as soon as i can)
one used an int[] ord index (the underlaying cache cannot be made per segment)
one used a cached DocIdSet created over the top level MultiReader (should be
able to have a DocIdSet per Segment reader here, but this will take some more
thinking (source of the matching docids is from a separate index), will also
need to know which sub docidset to use based on which IndexReader is passed to
scorer() - shouldn't be any big deal)
i'm a bit concerned that i may not be testing "multi-segment" searching quite
properly right now though since i think most of my indexes being tested only
have one segment.
On that topic, if i create a subclass of LogByteSizeMergePolicy and return null
from findMerges() and findMergesToExpungeDeletes() will this guarantee that
segments will only be merged if i explicitly optimize? In which case, i can
just pepper in some commits as i add documents to guarantee that i have more
than 1 segment.
Overall, i am really liking the per-segment stuff, and the Collector API in
general
its already made it possible to optimize a good deal of things away (like
calling Scorer.score() for docs that end up getting filtered away), however i
hit some deoptimization due to some of the crazy stuff i had to do to make
those 3 query implementations work, but this should only really be isolated to
one of the implementations (and i can hopefully reoptimize those cases anyway)
I would still like to see IndexSearcher passed to Weight.scorer(), and the
getIndexReaderBase() method added to IndexSearcher
This will clean up my current "hacks" to map docids
> Weight.scorer() not passed doc offset for "sub reader"
> ------------------------------------------------------
>
> Key: LUCENE-1821
> URL: https://issues.apache.org/jira/browse/LUCENE-1821
> Project: Lucene - Java
> Issue Type: Bug
> Components: Search
> Affects Versions: 2.9
> Reporter: Tim Smith
> Fix For: 2.9
>
> Attachments: LUCENE-1821.patch
>
>
> Now that searching is done on a per segment basis, there is no way for a
> Scorer to know the "actual" doc id for the document's it matches (only the
> relative doc offset into the segment)
> If using caches in your scorer that are based on the "entire" index (all
> segments), there is now no way to index into them properly from inside a
> Scorer because the scorer is not passed the needed offset to calculate the
> "real" docid
> suggest having Weight.scorer() method also take a integer for the doc offset
> Abstract Weight class should have a constructor that takes this offset as
> well as a method to get the offset
> All Weights that have "sub" weights must pass this offset down to created
> "sub" weights
> Details on workaround:
> In order to work around this, you must do the following:
> * Subclass IndexSearcher
> * Add "int getIndexReaderBase(IndexReader)" method to your subclass
> * during Weight creation, the Weight must hold onto a reference to the passed
> in Searcher (casted to your sub class)
> * during Scorer creation, the Scorer must be passed the result of
> YourSearcher.getIndexReaderBase(reader)
> * Scorer can now rebase any collected docids using this offset
> Example implementation of getIndexReaderBase():
> {code}
> // NOTE: more efficient implementation can be done if you cache the result if
> gatherSubReaders in your constructor
> public int getIndexReaderBase(IndexReader reader) {
> if (reader == getReader()) {
> return 0;
> } else {
> List readers = new ArrayList();
> gatherSubReaders(readers);
> Iterator iter = readers.iterator();
> int maxDoc = 0;
> while (iter.hasNext()) {
> IndexReader r = (IndexReader)iter.next();
> if (r == reader) {
> return maxDoc;
> }
> maxDoc += r.maxDoc();
> }
> }
> return -1; // reader not in searcher
> }
> {code}
> Notes:
> * This workaround makes it so you cannot serialize your custom Weight
> implementation
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
