-
Type:
Bug
-
Resolution: Duplicate
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
-
ALL
-
None
-
3
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Hello,
I have a large sharded collection with similar documents. A part of them have a field called "tag", which is always a string. I've indexed the field using partialFilterExpression:
{"v" : 2,
"key" : { "tag" : 1 },
"name" : "tag",
"ns" : "database.collection",
"partialFilterExpression" : {
"tag" : { "$exists" : true } }
}
However, when i run database.collection.count_documents({"tag": "value"}), the result takes a long time to complete. Running current_op shows me the query running:
"secs_running" : NumberLong(616), "microsecs_running" : NumberLong(616856589), "op" : "getmore", "ns" : "database.collection", "command" : { "getMore" : NumberLong("4801338963376724146"), "collection" : "collection", "planSummary" : "IXSCAN { tag: 1 }", "cursor" : { "cursorId" : NumberLong("4801338963376724146"), "createdDate" : ISODate("2019-11-26T13:23:57.281Z"), "lastAccessDate" : ISODate("2019-11-26T13:23:57.282Z"), "nDocsReturned" : NumberLong(0), "nBatchesReturned" : NumberLong(0), "noCursorTimeout" : false, "tailable" : false, "awaitData" : false, "originatingCommand" : { "$truncated" : "{ $truncated: \"{ aggregate: \"collection\", pipeline: [ { $match: { tag: \"value\" } }, { $group: { _id: { $const: 1 }, n: { $sum: { $const: 1 } } } } ], fro...\" }" },
So i'm running an aggregate, which i know can be slow. I remember running these kind of queries very fast, when the field was indexed. Is this because the index is a partialFilterExpression? If so, it looks like a bug, because i think it should be able to use the information from that index in order to complete a simple count operation.
I'm running mongod 4.2.1, and the count was issued with pymongo 3.9.0.
- duplicates
-
SERVER-28889 Partial index shouldn't do fetch on conditions that are true by the definition of the index
-
- Closed
-