[jira] [Updated] (LUCENE-9045) Do not use TreeMap/TreeSet in BlockTree and PerFieldPostingsFormat

["Drew Foulks (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/LUCENE-9045?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Drew Foulks updated LUCENE-9045:
--------------------------------
    Reporter: Bruno Roustant  (was: Bruno Roustant)

> Do not use TreeMap/TreeSet in BlockTree and PerFieldPostingsFormat
> ------------------------------------------------------------------
>
>                 Key: LUCENE-9045
>                 URL: https://issues.apache.org/jira/browse/LUCENE-9045
>             Project: Lucene - Core
>          Issue Type: Improvement
>            Reporter: Bruno Roustant
>            Assignee: Bruno Roustant
>            Priority: Major
>             Fix For: 8.4
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> TreeMap/TreeSet is a heavy structure designed to dynamically sort keys. It's iterator is much less performant than a list iterator. We should not use it when we don't need the sorting capability once built.
> And this is the case in BlockTreeTermsReader and PerFieldPostingsFormat. We need a Map and to sort keys at building time. But once built, we don't need to sort anymore, we can use a simple list for iteration efficiency.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org