Is the performance of $lookup still 130 times worse than Postgres?

Working with Data
,
1

Hello, I have a scenario where $lookup would be preferential to embedding and came across this article in researching its efficacy. https://www.enterprisedb.com/blog/comparison-joins-mongodb-vs-postgresql It states that because postgres is build with iterative substitution. as well as merge join and hash join that it is 130 times faster at joining with damning timespans for queries to run. My question is simple, is this still the case given it being 2 years old? I am aware of some recent improvements with $lookup including the ability to have either collection sharded now so I suppose I am hoping the performance of the stage itself has been improved.

EDIT: Set Slot-Based Query Execution seems to promise some performance improvement for $lookup but does it make a dent in the 130 times?

4

Landed here with the exact same question.
Guess that the answer is, “Yes, still” then, from the fact of the lack of replies.

5

From the mongo .local London event last week there has been a lot of improvement on the $lookup operator, like orders of magnitude performance boost.
I guess at the end of the day, try it with your workload, Mongo is not an relational database…if you need a relational database than choose one of them, if this type of operation is a key app path then you may need to model the documents differently.

I believe you may need to be running V7 to take advantage of all the slot processing changes, only some operations made use of it in V6