On 2010-08-06 19:33, Philippe Sigaud wrote:
2010/8/6 Adrian Matoga <e...@atari8.info <mailto:e...@atari8.info>>

    Hi,

    Is there any off the shelf solution for iterating over a range by
    chunks?


None that I know of.

    (should substitute [1, 2, 3], [4, 5, 6], [7, 8, 9], [10] for chunk
    in subsequent iterations)


As a data point, why do you think it should produce [10] and not stop at [7,8,9]?

By an analogy to taking a file by chunks. The other case also occured to me but I simply thought I could imagine more situations in which I need the whole input range to be exhausted, including remainder.

Here is what I cooked, it's still a bit rough around the edges. It has an optional step argument, to see how many elements to jump.

import std.range;
struct Chunks(R) if (isInputRange!R)
{
    R range;
    size_t n; // chunk size
    size_t step; // how many elements to jump

    bool empty() @property { return range.empty;}
ElementType!R[] front() @property { return array(take(range, n));} // inefficient if you call front() many times in a row
    void popFront() @property { popFrontN(range, step);}

    static if (hasLength!R)
        size_t length() @property { return (range.length+step-1)/step;}
}

Chunks!R chunks(R)(R range, size_t n) if (isInputRange!R)
{
    return Chunks!R(range, n, n); // default is step == n
}

Chunks!R chunks(R)(R range, size_t n, size_t step) if (isInputRange!R)
{
    return Chunks!R(range, n, step);
}

Thank you very much!

And yes, I also think it should be included in std.range (actually before posting the question I was almost sure I could find it there ;)).

Adrian

Reply via email to