Lucene query syntax and query_string

Sebastian_Stehle · November 9, 2021, 5:03pm

Hi,

I have issues to understand the Query_string operator.

I am building a few search providers for my CMS and I already implemented Elastic Search and Azure Cognitive Search. Because my main database is MongoDB, Atlas Search could be a good fit for the users, who actually use Atlas.

There are a few differences to azure and elastic and I don’t understand whether I am doing something wrong or not.

What all three service have in common is that they are built on top of lucene, therefore I though that the query_string operator uses the lucene query syntax as it seems to be done by elastic and azure: Apache Lucene - Query Parser Syntax.

Documentation about the elastic query string query can be found under: Query string query | Elasticsearch Guide [8.6] | Elastic

There are a few thing that work differently:

The Atlas query_string does not support fuzzy operator like helo~.
The Atlas query_string cannot search in multiple fields, because “*” is not supported as a default_path.
The Atlas query_string does not seem to support the lucene syntax at all.

Because Atlas search is also based on lucene I do not understand the reason for that. Is there something I do wrong? Is there a converter from lucene query syntax to the json structure from Mongo Atlas?

Sebastian_Stehle · November 9, 2021, 7:45pm

I have played around the lucene a little bit and I can probably make the parsing myself:

gist.github.com

https://gist.github.com/SebastianStehle/f8f3d27d99006903235f0c46eb2c349d

ParseQuery.cs

// See https://aka.ms/new-console-template for more information
using Lucene.Net.Analysis.Standard;
using Lucene.Net.Index;
using Lucene.Net.QueryParsers.Classic;
using Lucene.Net.Search;
using Lucene.Net.Util;
using MongoDB.Bson;
using MongoDB.Bson.IO;
using System.Globalization;
using System.Text;

This file has been truncated. show original

Have not tested it yet.