Does standard-folding analyzer in Neo4j perform stemming like the english analyzer?

Neo4j Graph Platform Procedures & APOC
, , , ,
1

I'm using Neo4j's full-text indexes and trying to decide between analyzers for indexing full names that may or may not include diacritics. I came across the standard-folding analyzer, and the documentation says:

"An analyzer that uses ASCIIFoldingFilter to remove accents (diacritics). Otherwise behaves as the standard English analyzer. Note! This analyzer may have unexpected behavior, such as tokenizing, for all non-ASCII numbers and symbols."

However, it's unclear whether "standard English analyzer" here refers to the english analyzer (which includes stemming and stop word filtering) or something else.

My Questions:

  1. Does the standard-folding analyzer also perform stemming like the english analyzer?
  2. If I want to index full names (e.g., "José García", "Jose Garcia"), where:
  • Diacritics should be ignored (e.g., "JosĂ©" = "Jose")
  • Stemming should NOT be applied
  • No stop-word filtering is needed→ What is the recommended analyzer to use in this case?

Thanks in advance for any clarification!

2

Hard to tell, I wrote a bit of code to test the analyzers. This is what standard-folding vs english generates in terms of tokens for "José García":

standard-folding=[jose, garcia]
english=[josé, garcía]

3

I am lazy, so if I have the room for it, I would save both a name and a normalized name property and use a normal range or text index. :) I guess it depends on the application....