How to correctly build indexes on MongoDB when every field will be searchable and sortable?

user_11063165 0
Viewed 562

I am designing a MongoDB collection that will have 50 million documents and every field in the document will be searchable and sortable. The searching and sorting logics will be sent from the frontend so could be a lot of fields searchings and sorting combinations. I've made some tests and concluded that when there is searching and sorting only in indexed fields the query runs very fast, but when searching or sorting non-indexed fields the query runs very slow.

Considering that will have a lot of possible searching/sorting combinations, how can I build indexes in this collection in this case to get a better performance?

mongoose aggregation-framework mongodb-query mongodb
1 Answers

Indexing comes at a cost of extra memory space and possible increased execution time of database write(insert and update) operations. However, like you rightly pointed out, indexing makes database reads(and sorting) super fast.

Creating indexes is easy and straight forward, however, you need to consider the tradeoffs, most times, this is usually the read-write ration of the fields in your documents.

If you frequently read(or sort) documents from a very large collection(like the 50million examples you mentioned), it makes a lot of sense to add indexing to all the fields you use to identify(or sort) your documents, you just need to ensure you don't run out of memory space in the DB. Not indexing the fields would be very frustrating, just imagine if you need to get the last document by a field that is not indexed, you would have to search through 49,999,999 documents to find it.

I hope this helps.

user_76858660
Related