Issue #25: Full Text Indexing and Query

Type: Feature RequestPriority: NormalStatus: ActiveReplies: 1

The ability of ObjectDB to index collections, including collections of strings, can be very useful in text indexing and search. By splitting document text into words and keeping these words in a List<String> persistent field, documents that contain specific words can be retrieved easily and efficiently.

This ability should be extended in future versions of ObjectDB to complete full text support, as provided by separate tools such as Apache Luncene.

ObjectDB Support

This feature would be useful as a built-in function.  But, it would also require the expansion of the SQL syntax to support other types of full-text operations like:

cat NEAR dog

whereby one can search for text where the token "cat" is within n characters of "dog".

Other types of things that would be useful would be fuzzy or synonym searches:


It would return records with "database", "datastore", "RDBMS", etc.


This would return "Smith", "Smyth", etc.

Another feature that would be really nice would be highlighting.  For example:

The red car is on the highway.

If I search for the token "red", I could get:

The <b>red</b> car is on the highway.

This would indicate that the bold word met the search criteria.