Feature #6499: Array::zip

• Status: Rejected
• Priority: Normal
• Assignee: matz (Yukihiro Matsumoto)
• Target version: 2.0.0

[prijutme4ty (Ilya Vorontsov)]

Sometimes it's more natural to have Array::zip rather than Array#zip. Otherwise first array of list became unnecessary dedicated
For example
cols = first_row.zip(second_row,third_row,fourth_row)
have more natural analog:
cols = Array.zip(first_row,second_row,third_row,fourth_row)

Implementation is obvious:
def Array.zip(first,*rest)
first.zip(*rest)
end

[mame (Yusuke Endoh)]

• Status changed from Open to Assigned
• Assignee set to matz (Yukihiro Matsumoto)

[trans (Thomas Sawyer)]

This is probably true of any Array method that take one or more other arrays as argument. Are there other's besides #zip?

[marcandre (Marc-Andre Lafortune)]

Hi

trans (Thomas Sawyer) wrote:

This is probably true of any Array method that take one or more other arrays as argument. Are there other's besides #zip?

I agree and would like Array.product to be considered as a worthy addition.

[duerst (Martin Dürst)]

The problem becomes even more obvious in the cases where one starts with an array of arrays:
New way:
Array.zip(*arrays)
Old way:
arrays[0].zip(*arrrays[1..-1])
The new way is clear and concise. The old way is a mess. I have definitely met this pattern for zip and product, maybe others. I always wished there had been a better way, but the idea I came up with was something like arrays.zips_many or so. That might work, too, but Array.zip is quite neat, too.

[duerst (Martin Dürst)]

duerst (Martin Dürst) wrote:

I always wished there had been a better way, but the idea I came up with was something like arrays.zips_many or so. That might work, too, but Array.zip is quite neat, too.

Sorry, it should have been arrays.zip_many above.

[prijutme4ty (Ilya Vorontsov)]

marcandre (Marc-Andre Lafortune) wrote:

Hi

trans (Thomas Sawyer) wrote:

This is probably true of any Array method that take one or more other arrays as argument. Are there other's besides #zip?

I agree and would like Array.product to be considered as a worthy addition.

Actually! I had such a problem with #product too, but have already forgotten about it.

[marcandre (Marc-Andre Lafortune)]

• Category set to core
• Target version set to 2.0.0

Wouldn't it be best if both Array.product and Array.zip were lazy?

Will someone produce a one-minute slide-show for this?

[claytrump (Clay Trump)]

• File array_product_zip.pdf array_product_zip.pdf added

I've had the same issue in the past, so here's a slide for this feature.

On Fri, Jun 1, 2012 at 2:54 AM, prijutme4ty (Ilya Vorontsov) <
[e-mail]> wrote:

Issue #6499 has been updated by prijutme4ty (Ilya Vorontsov).

Actually! I had such a problem with #product too, but have already
forgotten about it.

--

[mame (Yusuke Endoh)]

Received. Thank you!

--
Yusuke Endoh [e-mail]

[mame (Yusuke Endoh)]

• Status changed from Assigned to Rejected

Ilya Vorontsov and Clay Trump,

Sorry but this proposal was rejected at the developer meeting (7/21).

Here is a discussion summary:

• Matz was NOT positive to this kind of proposals that is for the
  sake of symmetric notation.

• Matz wants to save those who want Enumerator-version of Array#zip
  and product, but it does not require the new-style API.
  We should discuss it as another request.

• The behavior of Array.zip(*ary) is not trivial if ary == [].
  The spec candidates are:

  • returns an array whose size is the same as the first element, or
  • returns an array whose size is the same as the shortest element

If ary == [], both do not make sense because neither the first
element or the shortest one exist.
It may be intuitive to return an infinite array: [[], [], [], ...]

--
Yusuke Endoh [e-mail]

[trans (Thomas Sawyer)]

If it is any consolation, I will add these to Facets.

[marcandre (Marc-Andre Lafortune)]

I've thought a bit about this, and the proposed Array.zip basically already exists if all arrays have the same length.

It's called Array#transpose :-)

Array.zip([1,2,3], [4,5,6]) == [[1,2,3], [4,5,6]].transpose

Note that the case of Array.zip() could be debated, but, as explained in https://bugs.ruby-lang.org/issues/6852, [] == [].transpose is the best answer.

Because of this, it seems there is a lesser need for Array.zip.

On the other hand, I am still confident that Array.product or similar, would be helpful. I hope I won't offend anyone by opening a new request.

[marcandre (Marc-Andre Lafortune)]

• Is duplicate of Feature #8970: Array.zip and Array.product added