[Umbrella] std/collections

[kt3k]

We landed the initial version of std/collections at #993

This issue is the umbrella issue for the remaining 28 APIs, which still need discussion. More detailed type definitions / docs / examples are available in the original proposal PR by @LionC.

ARRAYS

Transformations to Array

• mapNotNullish Add mapNotNullish to collections #1103
• takeFirstWhile
• takeLastWhile
• dropFirstWhile
• dropLastWhile
• filterNotNullish
• sortBy Feature request: sortBy function in std/collections #1061
• runningReduce
• without feat(collections): Returns an array excluding single given value #1074
• withoutAll feat(collections): Add withoutAll that returns an array excluding all given values #1100

Transformations to value

• findLastIndex [Feature Request] Add findLastIndex function to std/collections #1060
• single
• firstNotNullishOf
• sumOf feat(collections): add sumOf #1108
• maxBy
• minBy
• maxWith
• minWith
• maxOf feat(collections): add maxOf #1144
• minOf

Number array transformations

• average moved to std/math proposal feat(statistics): Implement mean, median, and mode. #1111

Transformations to other formats

• associateBy Implement collections/associateBy #1071
• windowed

RECORDS

Predicates

• includesValue

Transformations to Records

• filterValuesNotNullish
• deepMerge feat(collections): add deepMerge #1075

Groups

• reduceGroups
• aggregateGroups

If you're interested in some specific items in the above, please pick one or more items, create a separate issue/PR for them, and keep discussion there.

[getspooky]

sounds good to me 👍🏼

[lowlighter]

I'm tempted to implement deepMerge but I saw some concerns about it in the mentioned discussion.
Is it still revelant ?

[grian32]

I would like to attempt implementing sumOf is there any issues with that particular function or am i good to go?

[LionC]

I would like to attempt implementing sumOf is there any issues with that particular function or am i good to go?

I would say just open the PR and if there are comments they will be posted there. Look at the ither functions for implementation and test reference, a lot can be copied as a template :-)

[majidsajadi]

do we really need average function? we can achieve this with a one-liner.

array.reduce((a, b) => a + b) / array.length

[LionC]

do we really need average function? we can achieve this with a one-liner.

array.reduce((a, b) => a + b) / array.length

There is a PR for a separate math mofule that contains average together with other functions.

A lot of things can be achieved with a one liner if you put enough things in a line, which does not necessarily mean that it is the most expressive, obvious, clear or often even known way to do it. See the math/statistics module PR for more info, average will not be in collections

[majidsajadi]

do we really need average function? we can achieve this with a one-liner.

array.reduce((a, b) => a + b) / array.length

There is a PR for a separate math mofule that contains average together with other functions.

A lot of things can be achieved with a one liner if you put enough things in a line, which does not necessarily mean that it is the most expressive, obvious, clear or often even known way to do it. See the math/statistics module PR for more info, average will not be in collections

I suggest that because we removed a lot of the proposed functions from the original PR. ref to this comment: #978 (comment)

[NotFounds]

I would like to implement single, is there any issue with this function?

[LionC]

I would like to implement single, is there any issue with this function?

I would just add the pointer that including (it) => true as a default for the predicate should be considered - that is what kt does because it allows a very smooth and readable transition from an (assumed) single-element array into that single element while also forcing you to handle what happens when that assumption is wrong:

// Assuming some cli that expects exactly one argument

const target = single(Deno.args)
    ?? error("Expected exactly one argument")

doMagic(target)

Also a note on the implementation: Make sure to short circuit when finding a second match instead of iterating through the rest of the array :-)

[kt3k]

The rest of items will be tracked in #1173.

Thanks @LionC for taking over the issue!