[
https://issues.apache.org/jira/browse/LUCENE-5833?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14149314#comment-14149314
]
Michael McCandless commented on LUCENE-5833:
--------------------------------------------
Thanks Varun.
Maybe instead of currentFieldsPosition name it nextFieldsPosition?
I.e, it's the index for the next value the iterator should return?
Maybe simplify this:
{noformat}
if (fieldVal.binaryValue() == null && fieldVal.stringValue() == null) {
continue;
}
return (fieldVal.stringValue() != null) ? new
BytesRef(fieldVal.stringValue()) : fieldVal.binaryValue();
{noformat}
to something like this:
{noformat}
if (fieldVal.binaryValue() != null) {
return fieldVal.binaryValue();
} else if (fieldValue.stringValue() != null) {
return new BytesRef(fieldValue.stringValue());
} else {
continue;
}
{noformat}
(And also where this pattern is used again towards the end of the
while loop). And I think also fix the similarly confusing
pre-existing payload null checking?
Can we move the {{BytesRef tempTerm}} type declaration down to where
we actually assign it?
Can you pull out the {{atLeast(N)}} from the for loops,
e.g. {{for(int j=0; j<atLeast(2); j++)}}? That is eval'd on each loop
iteration, so it's ... strange to call it over and over.
> Suggestor Version 2 doesn't support multiValued fields
> ------------------------------------------------------
>
> Key: LUCENE-5833
> URL: https://issues.apache.org/jira/browse/LUCENE-5833
> Project: Lucene - Core
> Issue Type: Bug
> Components: modules/other
> Affects Versions: 4.8.1
> Reporter: Greg Harris
> Assignee: Steve Rowe
> Attachments: LUCENE-5833.patch, LUCENE-5833.patch, SOLR-6210.patch
>
>
> So if you use a multiValued field in the new suggestor it will not pick up
> terms for any term after the first one. So it treats the first term as the
> only term it will make it's dictionary from.
> This is the suggestor I'm talking about:
> https://issues.apache.org/jira/browse/SOLR-5378
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
