About positional indexing

To enable faster phrase search performance and faster relevance ranking with the Phrase module, your project builds data files out of word positions. This process is called positional indexing.

The Oracle Endeca Server creates a set of positional data files for standard and managed attribute values.

Phrase search is automatically enabled in the Oracle Endeca Server at all times. For phrase search query processing, the Oracle Endeca Server examines potential matching text to verify the presence of the requested phrase query string. This examination process can be slow in the following cases:

The amount of text data is large (perhaps containing attribute values representing lengthy descriptions)
The text that is being processed is offline (in the case of document text)

The Oracle Endeca Server uses positional data files to improve performance in these scenarios. Positional indexing improves performance of multi-word phrase search, proximity search, and certain relevance ranking modules. The thesaurus uses phrase search, so positional indexing improves performance of multi-word thesaurus expansions as well. Positional indexing is enabled by default for attributes in your data store.