Best way to test predicate against a range

Jonathan M Davis jmdavisprog at
Sat Jun 26 19:10:37 PDT 2010

Okay. The functions in std.algorithm are quite powerful, but sometimes you have 
to play around with them a bit to figure out how to do exactly what you're trying 
to do. Sometimes it's a matter of figuring out exactly which algorithm really 
does what you want, and sometimes it's a matter of figuring out how to contort 
one of them to do what you want.

For example, there are two functions that I'd like to be have: all() and any(). 
That is, I want a function which checks a predicate against a range and returns 
whether all elements in that range satisfy the predicate, and I want a function 
that checks a predicate against a range and returns whether any element satisfies 
the predicate. Ideally, all() would shortcut if it found even one element which 
didn't satisfy the predicate, and any() would shortcut if it found even one that 

From the looks of it, canFind() is essentially any(), so that new addition to 
std.algorithm should deal with that. Previously, the best that I could figure out 
was to use find() for that and check whether the returned range was empty, which 
wasn't as efficient or elegant. So, that's a nice addition to phobos.

The problem is all(). I can't find any function which seems to effectively do 
all(). The best that I can think of is to use canFind() with a negated 
predicate, so if it returns true, it found an element which satisfied the 
negation of the predicate and therefore didn't satisfy the predicate. However, 
ideally, I'd be able to give a function the exact predicate that I'm looking for 
without having to contort the predicate to be able to use the function that I 
want. It feels like I'm dealing with only a || and I don't have a &&.

Is there a function in phobos which can't be used as an all(), and I'm just 
missing it, or is there at least a better solution than using canFind() with a 
negated predicate, or is canFind() with a negated predicate the best that there 
is at the moment?

- Jonathan M Davis

P.S. By the time that std.algorithm is completed, I suspect that it would be 
possible to write a small book on the myriad of ways to use it. It's extremely 
powerful and not really all that hard to use, but sometimes it takes a fair bit 
of thinking (for me at least) to figure out which function to use to do 
something. Hopefully I get better at it with time at least.

More information about the Digitalmars-d-learn mailing list