[vvpan]

Perhaps this is an opportunity to ask somebody who might know about BSON performance. As a POC/stress test for work I added two multi-GB datasets to Postgres (as JSONB) and to Mongo (BSON). While trying to query approximately a hundred megabytes of data (a few hundred documents) from each I found that Postgres executed the query and decoded the JSON data in under a second, while it took Mongo a few seconds. Does this mean that BSON is slow to deserialize? Or perhaps it is not related to serialization? I was quite confused.

[SahAssar]

Did you do a EXPLAIN ANALYZE on the postgresql query (and perhaps mongodb if it has something similar)?

It might help to find if it was the actual query or the de-serialization that was the bottleneck.