Get the latest tech news

Query Engines: Push vs. Pull (2021)

6 Apr 2021 People talk a lot about “pull” vs. “push” based query engines, and it’s pretty obvious what that means colloquially, but some of the details can be a bit hard to figure out.

For instance, Jamie Brandon has a post describing a general method for decorrelating subqueries in SQL that makes extensive use of DAG query plans in order to be efficient. As discussed in the Shaikhha paper: the merge join algorithm working is fundamentally based around the ability to traverse two iterators in lockstep, which is not practical in a push system where the consumer has little-to-no control. As noted in the Shaikhha paper, it’s hard to find apples-to-apples comparisons, since a lot of the migration to push models are motivated by a desire to compile queries to lower level code and the benefits that come from that cloud the results.

Get the Android app