How to count pymongo aggregation cursor without iterating

Question

I want to get the total number of records in an aggregate cursor in pymongo version 3.0+. Is there any way to get total count without iterating over the cursor?

cursor = db.collection.aggregate([{"$match": options},{"$group": {"_id": groupby,"count": {"$sum":1}}} ])
cursorlist = [c for c in cursor]
print len(cursorlist)

Is there any way to skip the above iteration?

In your above code len(list(cursor)) will give you the count of records returned. 2nd line is not needed therefore. — Sarath Nair, Nov 26 '15 at 08:56
Thanks @Sarath this will be useful if i don't want to update my aggregation query. — jadhavmahendra7, Nov 26 '15 at 13:12

score 7 · Answer 1 · edited Oct 03 '17 at 14:42

7

You could add another group pipeline where you specify an _id value of None to calculate accumulated values for all the input documents as a whole, this is where you can get the total count, as well as the original grouped counts, albeit in an accumulated array:

>>> pipeline = [
...     {"$match": options},
...     {"$group": {"_id": groupby, "count": {"$sum":1}}},
...     {"$group": {"_id": None, "total": {"$sum": 1}, "details":{"$push":{"groupby": "$_id", "count": "$count"}}}}
... ]
>>> list(db.collection.aggregate(pipeline))

edited Oct 03 '17 at 14:42

Vemonus

868
1
16
28

answered Nov 26 '15 at 08:46

chridam

100,957
23
236
235

This worked. Only minor change, Python have to use None and not null thanks! – jadhavmahendra7 Nov 26 '15 at 10:41
@jadhavmahendra7 Good catch! I've updated my answer to reflect this, cheers :) – chridam Nov 26 '15 at 10:43

How to count pymongo aggregation cursor without iterating

1 Answers1