Can’t find document at lucene index with no delimeter in phrase

I have document with text field fullName (the text is ‘1a‘). It saved to lucene index and indexed with custom analyzer (which contains WordDelimeterGraph filter with splitOnNumerics = 0 parameter).

The problem is that I can search that document only by query ‘1 a‘, not ‘1a’. Why?!

As I know WordDelimeterGraph filter (with splitOnNumerics = 0 parameter) must not(!) split literal and numeric to search the text ‘1a’ as is…

Indexing and searching are used the custom analyzer with the following filters:

  • whitespace tokenizer
  • lowercase filter (splitOnCaseChange=0, spltOnNumerics=0, protected=protected.txt)
  • stop filter (ignoreCaseTrue=0, format=snowball, words=words.txt)

Maybe anybody know the reason of it. I don’t have any ideas. Thanks!

P.S. I use lucene core version 9.0.0 (also reproduced on version 9.10.0).

Leave a Comment