AutomatonTermsEnum throws a NPE

Question

AutomatonTermsEnum throws a NPE

Closed this issue 8 years ago · 4 comments

This happens when the regexp has no trailing .*, even when it matches a value. e.g.:

http://localhost:8080/api/docvalues/dv_filename_set/ordered?filter=README

but add a .* and it works. Is this a Lucene bug?

! java.lang.NullPointerException: null
! at org.apache.lucene.index.AutomatonTermsEnum.(AutomatonTermsEnum.java:79)
! at com.github.flaxsearch.resources.DocValuesResource.getOrderedDocValues(DocValuesResource.java:174)

Answer 1 · 2017-01-06T16:07:14.000Z

Hm, this does look like a problem with lucene - AutomatonTermsEnum doesn't work with a fixed-string automaton.

As a workaround, maybe we should just always add a .* to the end of the passed-in filter, making it always be a prefix filter?

Answer 2 · 2017-01-11T11:52:58.000Z

I've opened a ticket in lucene: https://issues.apache.org/jira/browse/LUCENE-7627

Answer 3 · 2017-02-01T09:13:00.000Z

Is this fixed within the webapp, or just in the UI?

Answer 4 · 2017-02-01T11:04:02.000Z

It's fixed in the webapp. Try the example link.