Summary
Create a new index format that allows queries to distinguish whether the indexed field is missing or null without fetching the underlying document.
Motivation
The current index format forces us to choose between correctness and performance, for some queries. We can generate a correct plan with a FETCH stage, which has a performance cost; or, we can generate an efficient, covered plan that returns different results than a collection scan.
Cast of Characters
- Product Owner: Katya Kamenieva
- Project Lead: Pawel Terlecki
- Program Manager: Craig Homa
Documentation
Scope Document
Technical Design Document