However, this is also the nature of premature and micro-optimizations. The point of the benchmark was simply to show that assuming keys/values is slower than each is not always correct. Any decisions beyond that should only be measured based on actual code, with a profile, and an indication that this part of the code is the bottleneck.
In reply to Re^4: keys and values order on a hash
by Somni
in thread keys and values order on a hash
by citromatik
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |