We've migrated from a postgres database that was using vectors and using ## to_tsquery(:searchQuery) to perform queries. We are now using ElasticSearch and having a hard time getting the same accuracy that we saw with Postgres Vector searching.
For example we have the phrase "deadpool vs carnage" as the value for a title field on a document and would like this document to come back when performing a search with a query value of "dead carn".
When using postgres this wasn't a problem. But with ElasticSearch I'm unable to get this to return correctly.
I've tried different variations of all the match queries and can't get this to work. Any help would be very much appreciated
Elasticsearch is such a robust system that often times, like with many things, it's easy to overlook the simplest solution.
I ended up using the query_string search with wildcards in the string to achieve this.
You can pass in a query value of dead* carn*.
https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-query-string-query.html
Related
I am working on a website, where users can create entries into a database. These entries are all of the same form, so I was using Sqlite (with FTS5) so far, which is the only database I know ;)
However it is crucial that users can search these entries on the website. The full text search is working decently well (the users know approximately what they are looking for) but I need two improvements:
Wrong spelling should return the correct entry (I have seen the spellfix extension for sqlite for that, but I don't know how well it works)
more importantly if a user enters a few query words on the website I try to MATCH those with a sql query. If a user enters too many words, it will return 0 matches:
for example: if a user inputs "sars covid 19" into the search-bar:
CREATE VIRTUAL TABLE TEST USING FTS5(SomeText);
INSERT INTO TEST(SomeText) VALUES
('Covid 19');
SELECT SomeText
FROM TEST
WHERE SomeText MATCH 'sars covid 19';
=> 0 matches, but I would want it to return the 'covid 19' entry.
I realise that sqlite might be too simplistic for a database that needs to handle searches well. Now my question is: Do Postgres or MongoDB have search engines that include the functionality that I need? Or is it worth diving into solutions with Elastic Search etc.?
Most articles I found on this are 5-10 years old, so I am wondering what the current state of affairs is regarding search engines in databases. Any hints are greatly appreciated
Combination es + mongodb work well, you index and perform full text search in es and you keep the original documents with some key fields indexed in mongodb...
Elasticsearch will work for sure. You only have to think about how you will index your document, and you will be able to find them the way you index them, in your context it seems that the default text will work with a match query :
https://www.elastic.co/guide/en/elasticsearch/reference/current/full-text-queries.html
MongoDb will work too in this simple case : https://docs.mongodb.com/manual/text-search/ , but mongo wont work with tokenizer so if you need to upgrade your text search mongo will be limited.
Postgresql could do it, using the like but I am not familiar with enough, if you have 10k entries, it will be ok for sure, if you expect 1 millions, mongo or es would be better.
If you have to choose between mongodb and es, you have to be more specific in your question, for full text, es is really nice, having a lot of features, mongodb give some nice database tools too. Sometimes es will be better, sometimes mongo, depends of what you need. If you only want full text, es is a must.
I Want to query using part of id to get all the matched documents. So I tried “starts with” and "contains" which works find but is there any performance issue for large collection?
The best way to make this search optimum :
Add $text index on the fields you want to do search in. This is really important because internally it tokenize your string to that you could search for a part of it.
Use regex which is also quicker to do.
If you are using aggregate, read this mongodb official doc about aggregation optimization which might help you to implement this in efficient manner : https://docs.mongodb.com/manual/core/aggregation-pipeline-optimization/
Last but not the least, if you are not yet fully inclined towards mongodb and project is fresh, look out for elasticsearch service which is based on Lucene. Its extremely powerful doing these kinds of searches.
I am really new to the programming but I am studying it. I have one problem which I don't know how to solve.
I have collection of docs in mongoDB and I'm using Elasticsearch to query the fields. The problem is I want to store the output of search back in mongoDB but in different DB. I know that I have to create temporary DB which has to be updated with every search result. But how to do this? Or give me documentation to read so I could learn it. I will really appreciate your help!
Mongo does not natively support "temp" collections.
A typical thing to do here is to not actually write the entire results output to another DB since that would be utterly pointless since Elasticsearch does its own caching as such you don't need any layer over the top.
As well, due to IO concerns it is normally a bad idea to write say a result set of 10k records to Mongo or another DB.
There is a feature request for what you talk of: https://jira.mongodb.org/browse/SERVER-3215 but no planning as of yet.
Example
You could have a table of results.
Within this table you would have a doc that looks like:
{keywords: ['bok', 'mongodb']}
Each time you search and scroll through each result item you would write a row to this table populating the keywords field with keywords from that search result. This would be per search result per search result list per search. It would probably be best to just stream each search result to MongoDB as they come in. I have never programmed Python (though I wish to learn) so an example in pseudo:
var elastic_results = [{'elasticresult'}];
foreach(elastic_results as result){
//split down the phrases in this result and make a keywords array
db.results_collection.insert(array_formed_from_splitting_down_result); // Lets just lazy insert no need for batch or trying to shrink the amount of data to one go or whatever, lets just stream it in.
}
So as you go along your results you basically just mass insert as fast a possible create a sort of "stream" of input to MongoDB. It can do this quite well.
This should then give you a shardable list of words and language verbs to process things like MRs on and stuff to aggregate statistics about them.
Without knowing more and more about your scenario this is pretty much my best answer.
This does not use the temp table concept but instead makes your data permanent which is fine by the sounds of it since you wish to use Mongo as a storage engine for further tasks.
Actually there is MongoDB river plugin to work with Elasticsearch...
db.your_table.find().forEach(function(doc) { b.another_table.insert(doc); } );
I’m playing around with MongoDB for the moment to see what nice features it has. I’ve created a small test suite representing a simple blog system with posts, authors and comments, very basic.
I’ve experimented with a search function which uses the MongoRegEx class (PHP Driver), where I’m just searching through all post content and post titles after the sentence ‘lorem ipsum’ with case sensitive on “/I”.
My code looks like this:
$regex = new MongoRegEx('/lorem ipsum/i');
$query = array('post' => $regex, 'post_title' => $regex);
But I’m confused and stunned about what happens. I check every query for running time (set microtime before and after the query and get the time with 15 decimals).
For my first test I’ve added 110.000 blog documents and 5000 authors, everything randomly generated. When I do my search, it finds 6824 posts with the sentence “lorem ipsum” and it takes 0.000057935714722 seconds to do the search. And this is after I’ve reset the MongoDB service (using Windows) and this is without any index other than the default on _id.
MongoDB uses a B-tree index, which most definitely isn’t very efficient for a full text search. If I create an index on my post content attribute, the same query as above runs in 0.000150918960571, which funny enough is slower than without any index (slower with a factor of 0.000092983245849). Now this can happen for several reasons because it uses a B-tree cursor.
But I’ve tried to search for an explanation as to how it can query it so fast. I guess that it probably keeps everything in my RAM (I’ve got 4GB and the database is about 500MB). This is why I try to restart the mongodb service to get a full result.
Can anyone with experience with MongoDB help me understand what is going on with this kind of full text search with or without index and definitely without an inverted index?
Sincerely
- Mestika
I think you simply didn't iterate over the results? With just a find(), the driver will not send a query to the server. You need to fetch at least one result for that. I don't believe MongoDB is this fast, and I believe your error to be in your benchmark.
As a second thing, for regular expression search that is not anchored at the beginning of the field's value with an ^, no index is used at all. You should play with explain() to see what is actually happening.
We are trying to develop a strategy for using elasticsearch for full-text searching on our mongodb instance. It would appear that every key that we want to use as a filter must be included in elastics index. Potentially we could want to use every key in mongo as a filter - i.e. full-text search on description, filter by date and telephone number. Does anyone have any real-world experiences of adding full-text to mongo that they can share?
Maybe we can just use elasticsearch as a db?
I do not see any reason to use ElasticSearch in conjunction with MongoDb, just use ElasticSearch as separate document storage for documents, that have to be searched. And yes, you can even as whole db. Of course it depends on your domain model and other factors.
If you don't need stemming, fuzzy search, complicated wildcard search, you can do search with mongoDb. When new document inserted, split it to words in lower case, and add to the array "words" for example. Later you can perform search request against this array with regex. Not you can' use I (ignore case) option in this regex, and you can search only LIKE% wildcard (or without wildcard), otherwise search would not use mongoDb index.
One more option - you can try to find river for mongoDb
Another option - is to use Lucene if you are using Java. Probably you will be able to extend Directory class, in such a way, that Lucene will store index in MongoDb instead of file system or RAM. I have not made any research in this area, but I think it is possible
I experimented with full text search in MongoDB by splitting the words in the string like #Umar suggested. Honestly though, its a database and not a search engine so I would use Mongo for persistant storage and ElasticSearch for the search engine part of it. As a matter of fact, I would stick with something like Postgresql for persistant storage and then push the data you want to search out to the search engine. http://gdal.org/ogr/drv_elasticsearch.html is a driver that will allow you to quickly export your data from one RDBMS to ElasticSearch. THe data does not have to be geospatial in order to use it GDAL as long as their is a way to connect to the input source.
Adam