couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Constantin Teodorescu <>
Subject Re: Mango full text search is immune to accented letters?
Date Mon, 01 Aug 2016 17:34:34 GMT
On Mon, Aug 1, 2016 at 7:50 PM, Tony Sun <> wrote:

> Hey Teo,
>  Were you able to get Mango text search working? Specifying the analyzer
> get's a little tricky.

No Tony, thanks for asking!
I realised that there is another process that it handles Lucene search,
just like in the old "CouchDB Lucene"!
We have used "CouchDB Lucene" 3 years ago, with very, very good results but
I thought that CouchDB 2.0 included an Erlang port of Lucene, like Riak did
for full text search!

For the moment I decided to wait some more time for CouchDB 2.0 to
stabilise, including howto's documentation for installing! :-)

At the begining, I was somehow excited reading about Mango indexes.
Now I understood that "json" type indexes are basically views where map
functions are probably defined in Erlang.
Today I discovered that I cannot use "sort" in a query if the column is not
present in the index fields ... another deception! :-(
I believed that the query executor is using a Mango index then it is able
to do a merge sort from shards, like any SQL database is doing! :-D


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message