Re: Index of current iteration in for-each loop

date: Tue, 9 Jul 2024 16:57:42 -0400
from: John Yeung <gallium.arsenide@xxxxxxxxx>
subject: Re: Index of current iteration in for-each loop

Daniel, thanks for providing those insights! I do have a few
additional comments though:

On Tue, Jul 9, 2024 at 5:25?AM Daniel Gross <daniel@xxxxxxxx> wrote:

sorry - but I have to disappoint you. Every language that I know, that

has a "for-each" construct, assigns a copy of the actual array element to
the loop variable.

Indeed, that is the point of "for-each", and it's why there is even
terminology that is distinct from simply "for".

But Python, and perhaps a growing number of other languages, do make
it easy to get both the index and the element value in the same loop.

If you want to modify arrays, you have to use the indexed form of FOR:

for index = 1 to %elem(array);
...
end-for;

This is the classic idiom, and any time you need the array index, this
is the standard and usually best way to do it. (In most languages, not
just RPG).

Interesting question would be the question - does the RPG compiler and

runtime guarantee that for-each is iterating for 1 to %elem()?

That question would be interesting, because other languages do not give

any guarantee that "for-each" iterates the array elements in any particular
order.

What languages are you thinking of? I don't know of any that would
iterate *an array* out of order. There are some data structures that
are not intrinsically ordered (such as sets or hash maps), and using
the language-appropriate "for-each" over those might not follow the
desired order. RPG doesn't have any such data structures. Because of
the low-level notion of what an array is (contiguous, evenly spaced
addresses in memory), the simplest thing to do when implementing
"for-each" in the compiler is to iterate an array in index order.
Anything else would be simultaneously more complex to implement,
slower to execute, and less advantageous for program logic.

John Y.