Re: Formal equivalents using fold-left

> If we want to reduce the number of primitives that we depend on,
>then defining numbered-items() (for sequences), and numbered-members()
(for arrays) as primitives,
> using a recursive function for the formal equivalent, would seem to be
the best way forward.
> The question then is whether to make these user-visible functions, or to
use them for exposiition only,.

I find *fn:numbered-items *pretty nice and handy - will be glad to have it.
It is much better than repeatedly, again and again to zip a sequence with
(1 to 1 000 000), using *fn:for-each-pair *.

Thanks,
Dimitre.

On Wed, Mar 12, 2025 at 2:57 AM Michael Kay <mike@saxonica.com> wrote:

> Many functions with callbacks, such as filter() and for-each() have a
> formal equivalent using fold-left(), which in turn is defined in terms of
> recursion.
>
> Dropping the position argument from the fold-left callback means these
> need to be rewritten.
>
> In most case it's easy enough, for example fn:filter becomes
>
> fold-left($input ! {'item': ., 'pos': position()},
>           (),
>           fn($result, $pair) {
>             if ($predicate($pair?item, $pair?pos))
>             then ($result, $pair?item)
>             else $result
>           })
>
> There's the slight disadvantage that this introduces a dependency on "!"
> and "position()", which are themselves specified rather informally.
>
> Also, the same technique doesn't work for array functions like
> array:filter(), because we don't have a simple mapping operator for arrays.
>
> If we want to reduce the number of primitives that we depend on, then
> defining numbered-items() (for sequences), and numbered-members() (for
> arrays) as primitives, using a recursive function for the formal
> equivalent, would seem to be the best way forward. The question then is
> whether to make these user-visible functions, or to use them for
> exposiition only,.
>
> Michael Kay
>
>
>