Context Navigation

Changes between Initial Version and Version 1 of Ticket #35399, comment 4

initial	v1
1		Thank you for taking the time to perform the benchmarks, I think you missed the point brought up about hashing being expensive on non-literal values though.
	1	Thank you for taking the time to perform the benchmarks, I think you missed the point brought up about hashing being expensive (or simply not usable) on non-literal values though.
2	2
3	3	In other words `int.__hash__` is pretty fast but `Expression.__hash__` isn't and you need to perform an implicit one to gather values in `defaultdict(list)`. Since `bulk_update` support both expressions and literal assignment the benchmark must be run against both to be representative.