top button
Flag Notify
    Connect to us
      Site Registration

Site Registration

MongoDB: How to separate a large capacity of collection into a number of small files

0 votes
316 views

MONGODB 3.2.8, WIREDTIGER STORAGE ENGINE
SINGLE COLLECTION HAS MORE THAN 500G, WHETHER IT NEEDS TO BE SEPARATED?

WHAT PARAMETER SHOULD BE ADJUSTED ?

posted Oct 19, 2016 by anonymous

Looking for an answer?  Promote on:
Facebook Share Button Twitter Share Button LinkedIn Share Button

Similar Questions
+1 vote

I have a project, which has 1000 images. From each image, I get 1 million documents that I currently store in a mongoDB collection.

I have 2 alternatives,
1. Create a separate collection for each image, over multiple projects, say 10 projects, and each image(collection) would have 1 million documents that will be indexed. OR
2. Have a single collection and have 1000x1,000,000 documents in one collection, which again will be indexed.

My research and reading suggests that it is not a very good idea to create 1000 collections (for each image) and create an index for each one of 1000, and go with adding a billion documents to a single collection, to which mongoDB caters with no major issue.
My main aim is to enhance performance for reading. There would be no updates and deletes. Only reads and writes. I can afford writes to be slow, but not reads. So am I right in choosing the 2nd alternative?

Any guidance is highly appreciated.

+3 votes

Here I would like to explain my query a bit more.

I inserted multiple documents within a collection. Few documents have same set of keys and few of them have some extra key:value pairs . For example:
document 1 : key1 : "key1"
document 2: key1 : "key1", key2: "key2"
document 3: key1 : "key1", key2: "key2", key3 : "key3"
document 4: key1 : "key1", key3 : "key3"

Now I want to delete the documents which have key1 and key2.
What will be the command to do so ?

+1 vote

I have a DB with several (quite big) collections, on Mongo 3.2.5 and WiredTiger with snappy compression on. I would like to change the compression setting on one small collection I have to test it uncompressed. Is there a possibility to do that on a secondary without having to resync the whole DB?

...