[
https://issues.apache.org/jira/browse/LUCENENET-86?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12522192
]
Digy commented on LUCENENET-86:
-------------------------------
Hi George,
case SortField.FLOAT:
float f1 = (float)((System.Single)docA.fields[i]);
<------ Bug
float f2 = (float)((System.Single)docB.fields[i]);
1- Since the content of docA.fields[i] appears to be a string, it causes a
TypeCast exception. therefore i check its type in the patch and if it is a
string i convert it to float.
2- After fixing the above bug, the problem docA.fields[i]==null arises. this
needs another fix.
As you say, maybe it is a ".", vs. "," issue, coming out from somewhere else.
DIGY
Oooops,
***** After 5 minutes i wrote above message, i discovered that the sample data
in TestSort.cs contains floats with "."s not with ","s. Replacing dots with
commas solved the problem.
So, patch can wait, but i think i will be needed sometime.
DIGY
> Nunit test for TestSort.
> ------------------------
>
> Key: LUCENENET-86
> URL: https://issues.apache.org/jira/browse/LUCENENET-86
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Digy
> Priority: Minor
> Attachments: FieldDocSortedHitQueue.patch, Screenshot.JPG
>
>
> There are two bugs in FieldDocSortedHitQueue.LessThan method
> 1- in "case SortField.FLOAT:"
> The value of docA.fields[i] or docB.fields[i] may be a string. In
> this case SupportClass.Single.Parse should be invoked.
> 2- No null control for docA.fields[i] or docB.fields[i].
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
