[LUCENE-2136] MultiReader should not use PQ for its Term/sEnum if it has only 1 reader

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.0-ALPHA
• Component/s: core/index
• Lucene Fields: New
• Assignee: Michael McCandless
• Reporter: Michael McCandless
• Created: 08/Dec/09 10:49
• Updated: 28/Aug/22 12:16
• Resolved: 08/Dec/09 13:47

Description

Related to LUCENE-2130....

Even though we've switched to segment-based searching, there are still times when the Term/sEnum is used against the top-level reader. I think Solr does this, and from LUCENE-2130, certain rewrite modes of MTQ will do this as well.

Currently, on an optimized index, MTQ is still using a PQ to present the terms, which is silly because this just adds a sizable amount of overhead. In such cases we should simply delecate to the single segment.

Note that the single segment can have deletions, and we should still delegate. Ie, the index need not be optimized, just have a single segment.