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 2014/12/17 15:47:16 UTC
[jira] [Resolved] (UIMA-4155) use bulk operations, avoid sort/dedup
in serialization of indexed FS
[ https://issues.apache.org/jira/browse/UIMA-4155?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Marshall Schor resolved UIMA-4155.
----------------------------------
Resolution: Fixed
> use bulk operations, avoid sort/dedup in serialization of indexed FS
> --------------------------------------------------------------------
>
> Key: UIMA-4155
> URL: https://issues.apache.org/jira/browse/UIMA-4155
> Project: UIMA
> Issue Type: Improvement
> Components: Core Java Framework
> Reporter: Marshall Schor
> Assignee: Marshall Schor
> Priority: Trivial
> Fix For: 2.7.0SDK
>
>
> The various serialization methods which start from the indexed types call the getIndexedFSs to get the indexes for 1 view. This method does extra array creation and sorting / deduping, not needed if there's a guarantee of no duplicates. This is now possible due to the change to add default bag indices (no need to merge "set" indices), and if allow_duplicate_add_to_indices is not specified. Eliminate this extra overhead in that case, and also use "bulk" methods where possible that avoid iterating and use things like System.arraycopy to accumulate the results.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)