-
Type:
Improvement
-
Resolution: Unresolved
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: Performance
-
None
Context
Currently, inflating (decoding to a dict) a RawBSONDocument only uses the C extensions to decode each element, with the dict-building loop running in pure Python. Calling from Python to C for every single element is significantly slower than calling into C once for the entire document.
Definition of Done
- We should match the same parsing pattern for other BSON optimizations in C and make RawBSONDocument match
Pitfalls
- N/A