CODEWITHSUNDEEP
mongodbmongodb-query
HC1122
HC1122

Reputation: 414

How to speed up getting result from large collection?

I'm working with MongoDB and I want to display last 10 entries in collection, which should be ordered by score. The problem is that there is about 2 million entries so each query takes up from 5-7 sec.

Things were working pretty good when the database was smaller but with increasing size problems occurred. I checked .explain('executionStats') and I noticed that in result there is nReturened: 10 and totalDocsExamined: 1959933 which means that it is going through whole database to return last 10 entries. (also totalKeysExamined: 0 if someone needs that info)

Searching query looks like this: db.logs.find({}).sort({"score":-1}).limit(10).toArray() Stages of query are COLLSCAN, SORT_KEY_GENERATOR, SORT

I'm aware that find({}) is going through whole DB, then sorting it, and getting last 10 results. But what should be the better way to speed the things up?

Upvotes: 1

Views: 1284

Answers (2)

simagix
simagix

Reputation: 1922

Create an index on {score: 1} will help greatly. Mongo definitely uses index for sorting if available. You should see IXSCAN, FETCH, and LIMIT stages after creating the index.

Upvotes: 3

Siddharth Hudda
Siddharth Hudda

Reputation: 66

You need to create index on key you are performing search in mongodb. Indexing will help you to retrieve data faster.

Use: db.collection.createIndex(keys, options)

For more information please read the doc MongoDB create index

Hope this will help. Good luck!

Upvotes: 1

Related Questions