You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Marshall Schor (JIRA)" <de...@uima.apache.org> on 2015/05/05 05:43:05 UTC

[jira] [Resolved] (UIMA-4282) Speeding up set iterators

     [ https://issues.apache.org/jira/browse/UIMA-4282?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Marshall Schor resolved UIMA-4282.
----------------------------------
    Resolution: Fixed

> Speeding up set iterators
> -------------------------
>
>                 Key: UIMA-4282
>                 URL: https://issues.apache.org/jira/browse/UIMA-4282
>             Project: UIMA
>          Issue Type: Improvement
>          Components: Core Java Framework
>            Reporter: Marshall Schor
>            Priority: Minor
>
> It is now trivially possible to speed up UIMA Set iterators which are iterating over a Type have subtypes, by changing the interator from one which is constantly computing the "next" element by looking at the sort order the current iterator and other iterators over other (sub)Types, to one which simply returns all the elements of 1 (sub)type, and then goes on to the next (sub)type.
> This would be a behavior change, in that elements returned would no longer be in a sort order.  However Set Indexes are documented to not have an "order" (see section 2.4.1.5 in the UIMA Reference, where it says:  "Set indexes do not enforce ordering, but use defined keys to specify equivalnce classes; addToIndexes will not add a Feature Structure to a set index if its keys match those of an entry already in the index."
> If users accidentally rely on sorted ordering among a Type and its subtypes for sets, then this change will violate that constraint.  
> If many people use this current property of UIMA Sets, it may not be a good thing to do this change.   



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)