NoSQL-like repositories like Lucene/Solr are generally more flexibles
and are tailored to work with full-text, however if you need to
perform complex relational operations or need to manage trasactions
then a DB is best.
On 5/29/11, Otis Gospodnetic wrote:

I'm guessing both Lucene and a DB (relational or not) may be about the same
here. Query like name="John" and age=30 and city="London" could be done
either, but if you think you'll need to expand those queries to include
full-text search, then I'd go with Lucene (or Solr).

Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Lucene ecosystem search :: http://search-lucene.com/

----- Original Message ----
From: "wave@vancameron.net" <wave@vancameron.net>
To: general@lucene.apache.org
Sent: Sat, May 28, 2011 2:28:00 AM
Subject: Query performance, bitset, btree

Hello, I'm new to Lucene. I have a question about performance. I have a
structured dataset split up into fields, for example name, age, city,
I want to query this dataset for exact matches to fields, e.g.
and age=30 and city="London". For such a dataset, how will performance of
Lucene compare with a database table with an appropriate index (e.g.
on name, age, city)? The dataset is pretty static so index update
performance is not an issue for me.

From what I understand, Lucene uses bitset indexes while a typical
index will use a b-tree (or hash). I'll admit I don't fully understand
a bitset index is. What types of datasets and queries take best advantage
bitset indexes vs a btree index?

View this message in context:

Sent from the Lucene - General mailing list archive at Nabble.com.
Sent from my mobile device

Search Discussions

Discussion Posts


Related Discussions

Discussion Navigation
viewthread | post
posts ‹ prev | 3 of 3 | next ›



site design / logo © 2016 Grokbase