Re: how do I efficiently remove one hash from another?

[davido]

in reply to how do I efficiently remove one hash from another?

Whether the loop is explicit, or implicit, there's a loop. But one cool means is:

delete @hash1{ keys %hash2 };

...which is pretty much the same thing as...

delete $hash1{$_} for keys %hash2;

...but with an implicit loop (via the hash slice) rather than the explicit 'for' loop. Note in either case, there's no need to worry about checking exists: delete doesn't complain if the element doesn't already exist.

[perltux]

Thanks for the detailed explanation.

[tobyink]

The former compiles to a much smaller op tree than the latter:

perl -MO=Concise -e'delete @hash1{keys %hash2}'
perl -MO=Concise -e'delete $hash1{$_} for keys %hash2'

On my machine, the slice performs about 20% faster than the loop.

[ColonelPanic]

The slice option only calls delete once. It also doesn't have to go the trouble of assigning $_ for each element. So that makes sense.