SolrInputDocument should preserve Field addition order
------------------------------------------------------
Key: SOLR-439
URL: https://issues.apache.org/jira/browse/SOLR-439
Project: Solr
Issue Type: Improvement
Components: clients - java
Reporter: Grant Ingersoll
Priority: Minor
The SolrInputDocument currently uses a Map for storing the names of Fields.
While in general this is fine, there are cases where it is important to
preserve the ordering of Fields as Lucene does. For instance, some
applications may only need to return certain fields w/ the results, in which
case some optimization could be attained by having them occur first in the
document so that Lucene doesn't have to skip over other Fields (if using the
FieldSelector to restrict what Fields get loaded)
Also, in some cases it is possible to setup Analyzers that coordinate results
between Fields by using the TeeTokenFilter (see LUCENE-1058 and
http://www.gossamer-threads.com/lists/lucene/java-dev/55927) thus saving on
indexing time.
Simply changing line 42 of SolrInputDocument to:
_fields = new LinkedHashMap<String,SolrInputField>();
should do the trick. This requires no interface changes, etc. and should be
back-compatible.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
