Hibernate Search: Using TermQuery on CSV values

Film Noir · **Joined:** Fri Oct 10, 2008 9:17 am **Posts:** 13

Greetings,

Currently we have a field stored in the index as comma separated values, so the field looks something like this:

07a6d7137fc3c60e035011619b8e652c,e92ae301f19f0f0680e685ae91dfa6b2,ab5deecf01de4bb95f9717420babe25b

We want to be able to perform an exact match search on one of the values, like this:

Code:

BooleanQuery query = new BooleanQuery();
query.add(new TermQuery(new Term("ids", "e92ae301f19f0f0680e685ae91dfa6b2" )), BooleanClause.Occur.MUST);

..etc

But this will not return any results despite the value existing in the field.

This is how the field is currently mapped:

Code:

@Field(index = org.hibernate.search.annotations.Index.TOKENIZED,store=Store.YES)
@FieldBridge(impl=CollectionToCSVBridge.class)
private Set<String> ids;

Any thoughts on how we can achieve this?

sanne.grinovero · **Posted:** Sat Jun 13, 2009 3:59 am

Hi,
you have to consider what exactly is going to be inserted in your index; in the case of

Quote:

07a6d7137fc3c60e035011619b8e652c,e92ae301f19f0f0680e685ae91dfa6b2,ab5deecf01de4bb95f9717420babe25b

you want to make sure this String gets split at the appropriate points (the commas).

The most efficient solution is to use an Analyzer with a custom Tokenizer; it's the Tokenizer which will transform your free text in a sequence of tokens. I'd suggest to use the Luke tool to look into your index to verify how the text gets tokenized.

On the reference documentation you can find some examples of defining an Analyzer using a custom Tokenizer.