bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Marco Wahl]

I suggest adding seq-set-subset-p to seq.el to realize the typical math
subset predicate.

(cl-defgeneric seq-set-subset-p (sequence1 sequence2 &optional testfn)
  "Return non-nil if the elements of SEQUENCE1 are contained in SEQUENCE2.
This does not depend on the order of the elements.
Equality is defined by TESTFN if non-nil or by `equal' if nil."
  (seq-every-p (lambda (item1) (seq-contains-p sequence2 item1 testfn)) sequence1))

;; test:
(should (seq-set-subset-p "ab" "abc"))
(should (seq-set-subset-p "ba" "abc"))
(should-not (seq-set-subset-p "dabc" "abc"))


That functionality is missing AFAICS.

The function could be placed naturally between the functions `seq-contains-p'
and `seq-set-equal-p'.

BTW function `seq-set-equal-p' could be rewritten using
`seq-set-subset-p'.


Thanks for reading,
-- 
Marco

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Juri Linkov]

> BTW function `seq-set-equal-p' could be rewritten using
> `seq-set-subset-p'.

Isn't the word "set" redundant in the name?
In `seq-set-equal-p' the word "set" is needed
to say "is set equal".  But in `seq-set-subset-p'
the word "set" doesn't add much meaning, so better
would be to name your new function just `seq-subset-p'.

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Marco Wahl]

Juri Linkov <juri@linkov.net> writes:

>> BTW function `seq-set-equal-p' could be rewritten using
>> `seq-set-subset-p'.
>
> Isn't the word "set" redundant in the name?
> In `seq-set-equal-p' the word "set" is needed
> to say "is set equal".  But in `seq-set-subset-p'
> the word "set" doesn't add much meaning, so better
> would be to name your new function just `seq-subset-p'.

Agreed.

The motivation for the suggested name `seq-set-subset-p' was to stay
close to the name `seq-set-equal-p'.  A user could interpret the prefix
"seq-set" as hint that the function implements a set operation.


Thanks!

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Juri Linkov]

>>> BTW function `seq-set-equal-p' could be rewritten using
>>> `seq-set-subset-p'.
>>
>> Isn't the word "set" redundant in the name?
>> In `seq-set-equal-p' the word "set" is needed
>> to say "is set equal".  But in `seq-set-subset-p'
>> the word "set" doesn't add much meaning, so better
>> would be to name your new function just `seq-subset-p'.
>
> Agreed.
>
> The motivation for the suggested name `seq-set-subset-p' was to stay
> close to the name `seq-set-equal-p'.  A user could interpret the prefix
> "seq-set" as hint that the function implements a set operation.

I guess the only reason why `seq-set-equal-p' has the prefix `seq-set-'
is because it would be too ambiguous without the prefix: `seq-equal-p'.

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Marco Wahl]

Juri Linkov <juri@linkov.net> writes:

>> The motivation for the suggested name `seq-set-subset-p' was to stay
>> close to the name `seq-set-equal-p'.  A user could interpret the prefix
>> "seq-set" as hint that the function implements a set operation.
>
> I guess the only reason why `seq-set-equal-p' has the prefix `seq-set-'
> is because it would be too ambiguous without the prefix: `seq-equal-p'.

I guess so too.

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Michael Heerdegen]

Marco Wahl <marcowahlsoft@gmail.com> writes:

> I suggest adding seq-set-subset-p to seq.el to realize the typical
> math subset predicate.

If I recall correctly, Stefan wanted a set.el library.  I only want to
mention that, because... well, it's related (and we already found that
"subsequence" testing is not what this implements).

Michael.

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Lars Ingebrigtsen]

Michael Heerdegen <michael_heerdegen@web.de> writes:

> If I recall correctly, Stefan wanted a set.el library.  I only want to
> mention that, because... well, it's related (and we already found that
> "subsequence" testing is not what this implements).

Yes, I'd rather have a set.el than adding some set commands to seq.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Juri Linkov]

>> If I recall correctly, Stefan wanted a set.el library.  I only want to
>> mention that, because... well, it's related (and we already found that
>> "subsequence" testing is not what this implements).
>
> Yes, I'd rather have a set.el than adding some set commands to seq.

There are already some commands in seq that look like set commands:
seq-union, seq-intersection, seq-difference.

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Michael Heerdegen]

Juri Linkov <juri@linkov.net> writes:

> There are already some commands in seq that look like set commands:
> seq-union, seq-intersection, seq-difference.

Yes, but that's unfortunate, and we need to stop at some point.  These
functions also don't really fit in seq.el.

Michael.

bug#52383: 29.0.50; [suggestion] Add seq-set-subset-p to seq.el

[Lars Ingebrigtsen]

Michael Heerdegen <michael_heerdegen@web.de> writes:

>> There are already some commands in seq that look like set commands:
>> seq-union, seq-intersection, seq-difference.
>
> Yes, but that's unfortunate, and we need to stop at some point.  These
> functions also don't really fit in seq.el.

And they're too slow to be used for even smallish sets.  So we should
grow a set.el library that is more efficient CPU-wise.  (The set
functions in seq.el are efficient RAM-wise, though.)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no