MongoByte MongoDB Logo

Welcome to the new MongoDB Feedback Portal!

{Improvement: "Your idea"}
We’ve upgraded our system to better capture and act on your feedback.
Your feedback is meaningful and helps us build better products.

Add new feedback Subscribe
Status Completed
Categories Atlas Search
Created by Guest
Created on Jun 16, 2020

RELATED FEEDBACK

Support Partial Indexing

We have a large collection but would only like to index the most recent entries, which would reduce the amount of space and resources it would take to index and query.
  • ADMIN RESPONSE
    Oct 18, 2025
    You can now create search indexes on a filtered collection defined by a $match stage with a $expr (Public Preview). See docs here ( https://www.mongodb.com/docs/atlas/atlas-search/transform-documents-collections/ ).
  • Guest
    Mar 28, 2025
    @Amy it's Public Preview and it's not working in local environment (using the mongodb/mongodb-atlas-local docker image), because setting up search index on view works only in Atlas UI. When does the support for mongosh, Compass and Drivers be available?
    Reply
  • Guest
    Mar 17, 2025
    @Martin you can do partial indexing if you use a view (with a $match stage) for your search index: https://www.mongodb.com/docs/atlas/atlas-search/transform-documents-collections/
    Reply
  • Guest
    Mar 17, 2025
    +1. This feature would help us too!
    Reply
  • Guest
    Oct 25, 2024
    Our application uses the Mongo Atlas search engine for full-text searching. The application stores documents in a single collection, indexed by an arbitrary number of Atlas search indexes corresponding to the number of tenants. The document contains some common fields and each tenant can enrich the documents with exclusive/unique fields/attributes. Since each index indexes all the documents of a collection, this becomes very resource-intensive regarding indexing time, RAM, and disk. At the moment, the only workaround that works for us is to move the tenants to their own collections, but this will make it difficult to access the whole set of documents.
    Reply
  • Guest
    Oct 25, 2024
    +1. This feature would help us too!
    Reply
  • Guest
    Oct 25, 2024
    Critical for our needs
    Reply
  • Guest
    Apr 11, 2024
    Would be nice, we would like to ignore soft-deleted records
    Reply
  • Guest
    Sep 12, 2022
    Upvote. Highly needed, as we have to index circa 60% of our documents, it can save a lot of hardware resources
    Reply
  • Guest
    Jul 1, 2022
    Being able to keep non searchable documents out of the index is very important. helps - keep secure documents more secure by helping to avoid accidental query errors which leak protected content - improves search performance
    Reply
  • Guest
    Nov 24, 2021
    Sooooooooooo important!!!!!!!!!!!!!!!!!
    Reply
  • Guest
    Nov 11, 2021
    This would be strongly needed. The current workaround is to create a separate collection, but then we cannot use a $out stage to populate it without destroying the index as said here: https://docs.atlas.mongodb.com/reference/atlas-search/create-index/ We are currently solving with a $merge stage but a partial index would be definitely a better solution.
    Reply
  • Guest
    May 11, 2021
    We dont want this data to be searchable with mongo search, but searchable with mongodb basic indexing. Mongo Search index performance will be much better if we can have this option and keep old data to be retriever with mongot index.
    Reply
  • Guest
    May 11, 2021
    Today, you may be able to move documents in a collection that you do not want searchable to online archive.
    Reply
  • Guest
    May 6, 2021
    Any news about this feature ? We have more than 100 millions documents and only 600k public searchable. This can improuve the performance without upgrading server.
    Reply
  • Guest
    Apr 12, 2021
    In several collections, we have tens of millions documents, but only 1-5 million are public/searchable in any given collection. It seems tragic to index all 50 million documents when only 5 million are public.
    Reply