Just pushed a big feature to ES called Percolator. Percolator does the "reverse" of what ES does today. Instead of indexing docs and then running searches against them, one registers queries and percolates a doc among them, getting back which queries matched that doc.
Tons of more features can be built on top of this core feature, would love to hear some feedback.
Loving the pace of innovation! This looks really cool, exactly the
type of idea Paul and I were thinking about when we asked about email
alerts.
Would this work for a query which spans multiple indexes?
We're set to upgrade our production instances next week from 0.13 to
0.14.2 to take advantage of the substantial reduction in disk space
needed for full text queries w/ highlighting.
Would this work for a query which spans multiple indexes?
It works a bit differently then the idea of searching across indices. You register a query under an index, and then you can percolate a doc against those queries.
If you have several indices, you will need to register queries on each one.
On Thursday, January 13, 2011 at 6:00 PM, dbenson wrote:
Shay,
Loving the pace of innovation! This looks really cool, exactly the
type of idea Paul and I were thinking about when we asked about email
alerts.
Would this work for a query which spans multiple indexes?
We're set to upgrade our production instances next week from 0.13 to
0.14.2 to take advantage of the substantial reduction in disk space
needed for full text queries w/ highlighting.
Just pushed a big feature to ES called Percolator. Percolator does the
"reverse" of what ES does today. Instead of indexing docs and then running
searches against them, one registers queries and percolates a doc among
them, getting back which queries matched that doc.
Tons of more features can be built on top of this core feature, would
love to hear some feedback.
Note, this feature is still experimental.
Issue: https://github.com/elasticsearch/elasticsearch/issues/issue/624
How would perculator play together with something like the twitter
river? Thinking to make this really useful you'd want some kind of
callback / webhook - http://wiki.webhooks.org/ - to be fired when
registered queries match. Or am I missing something?
Just pushed a big feature to ES called Percolator. Percolator does the "reverse" of what ES does today. Instead of indexing docs and then running searches against them, one registers queries and percolates a doc among them, getting back which queries matched that doc.
Tons of more features can be built on top of this core feature, would love to hear some feedback.
This feature is the core feature to get things going. Doing things like getting matched queries when indexing docs, and possibly publishing them is going to be built on top of it.
On Friday, January 14, 2011 at 1:29 PM, harryf wrote:
How would perculator play together with something like the twitter
river? Thinking to make this really useful you'd want some kind of
callback / webhook - http://wiki.webhooks.org/ - to be fired when
registered queries match. Or am I missing something?
Just pushed a big feature to ES called Percolator. Percolator does the "reverse" of what ES does today. Instead of indexing docs and then running searches against them, one registers queries and percolates a doc among them, getting back which queries matched that doc.
Tons of more features can be built on top of this core feature, would love to hear some feedback.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.