cost of collection.count?

[Maciej Dziardziel]

I'd like to know if getting amount of documents in collection
(collection.count()) is expensive
on large colelctions? In otherr words: does mongo keep tracks of this
number and can provide it instantly or does it perform something like
full scan or other calculations that can be slow when there is lot of
documents?

--
Maciej Dziardziel

[Eliot Horowitz]

Count with no filter is fast, we keep track of the number of objects
in a collection
If you specify a query, then it has to so the query

On Nov 27, 2009, at 7:52 AM, Maciej Dziardziel <fie...@gmail.com>
wrote:

> --
>
> You received this message because you are subscribed to the Google
> Groups "mongodb-user" group.
> To post to this group, send email to mongod...@googlegroups.com.
> To unsubscribe from this group, send email to mongodb-user...@googlegroups.com
> .
> For more options, visit this group at http://groups.google.com/group/mongodb-user?hl=en
> .
>
>

[Dwight Merriman]

yes it is instantaneous on large collections on a single mongod server

with sharding, it will be reasonably fast but get a bit slower as the collection grows

[Maciej Dziardziel]

Why and how is it different in case of sharding?

> with sharding, it will be reasonably fast but get a bit slower as the
> collection grows

--
Maciej Dziardziel

[Eliot Horowitz]

Because we have count each chunk separately to prevent miscounts
during migrations, etc...
Otherwise the number could be wrong at a given moment.