mongodb - look up table like SQL?

Question

Suppose I have a data structure in the form of [String, int], I would like to be able to do lookups in O(1) time. In traditional SQL databases, I can simply create a look-up table.

How can I translate the same to mongodb such that I can use the String element as the key and int as the value?

Answer 1

In MongoDB, your documents would need to be structured like this:

{
   key: "the key name"
   value: 100
}

Then, you'd create an index:

db.myLookupTable.ensureIndex({key : 1})

It's not going to be O(1) time though as it's not a Hash lookup. As explained here, indexes, by default, in MongoDB are B-tree based so that other types of operations are possible (like sorting which you could not get from a Hash-based index). B-trees are worst case O(log n).

You could also enable the Hashed indexes for an index in MongoDB:

db.myLookupTable.ensureIndex({key: "hashed" })

That limits the types of queries you can do on the index though. For example, range queries are not possible.

Answer 2

MongoDB allows you to create an index on fields in a collection; however, these are B-tree indexes and would result in lookups that are O(log n).

If you need O(1) lookups, then I would suggest using Redis.