The Inbox: Collections-mt.832.mcz
Locked
 
|
A new version of Collections was added to project The Inbox:
http://source.squeak.org/inbox/Collections-mt.832.mcz ==================== Summary ==================== Name: Collections-mt.832 Author: mt Time: 9 May 2019, 3:27:06.069782 pm UUID: b3c087f4-c2bc-d542-91d7-9c214ae3a97f Ancestors: Collections-nice.831 Adds #take: to create a sub-collection from any collection by specifying the number of elements. Works like #first: for sequenceable collections but does not fail if collection is too small. =============== Diff against Collections-nice.831 =============== Item was added: + ----- Method: CharacterSet>>take: (in category 'accessing') ----- + take: n + + self shouldNotImplement.! Item was added: + ----- Method: Collection>>take: (in category 'accessing') ----- + take: n + "Enumerate this collection and return the first n elements or less." + + | index result | + index := 1. + result := self species new: (n min: self size). + self associationsDo: [:each | + result add: each. + (index := index + 1) > n ifTrue: [^ result]]. + ^ result! Item was added: + ----- Method: SequenceableCollection>>take: (in category 'accessing') ----- + take: n + + ^ self first: (n min: self size)! Item was added: + ----- Method: Stream>>take: (in category 'accessing') ----- + take: n + + ^ self next: n! |
Re: The Inbox: Collections-mt.832.mcz
|
|
I note false polymorphism with Integer>>take: of course both are related, since Integer>>take: (11 take: 4) tells how many different ways you can #take: 4 elements from a Set of size 11... Le jeu. 9 mai 2019 à 15:27, <[hidden email]> a écrit : A new version of Collections was added to project The Inbox: Why? a simple do: loop seems feasible.
+ ^ result! First reaction: I did not understand associationsDo:... Why not just do:? After verification: I see that default Behavior of associationsDo: is to do: and this avoids a redefinition of take: in Dictionary... I don't find it unsurprising, but if it works...
|
Re: The Inbox: Collections-mt.832.mcz
|
Hmmm... (Set withAll: (1 to: 11)) take: 4 "-> e.g. #(1 2 3 4)" (Set withAll: (1 to: 11)) size take: 4 "-> 330" Well, the difference is that Integer >> #take: considers duplicate elements,too. Anyway, could such false polymorphism be a problem in a way that it would be hard to debug? Best, Marcel
|
Re: The Inbox: Collections-mt.832.mcz
|
|
Le ven. 10 mai 2019 à 08:04, Marcel Taeumel <[hidden email]> a écrit :
| count | count := 0. (1 to: 11) combinations: 4 atATimeDo: [:p | (p asSet size = p size) ifFalse: [self halt: 'duplicate']. count := count + 1]. ^count
I don't think so.
|
|
|
In reply to this post by marcel.taeumel
On 10/05/19 11:34 AM, Marcel Taeumel wrote:
> Hmmm... > > (Set withAll: (1 to: 11)) take: 4 "-> e.g. #(1 2 3 4)" > (Set withAll: (1 to: 11)) size take: 4 "-> 330" May I suggest using #slice: or #copyFirst: for collections? slice: 10 copyFirst: 10 Regards .. Subbu |
Re: The Inbox: Collections-mt.832.mcz
|
|
I think that "first" in the name would not work for an arbitrary collection, which might not have any order at all. "copy" sounds too low-level. I do like #slice:. The name #take: is inspired from libraries for other languages and database frameworks. Yet, #slice: might also be better for randomized sampling (if desired) than #take:. Then, we have #anyOne. So, what about #some:? ... or #anySome:? :-D Best, Marcel
|
Re: The Inbox: Collections-mt.832.mcz
|
|
Le ven. 10 mai 2019 à 10:25, Marcel Taeumel <[hidden email]> a écrit :
Agree, we already have #first: for the Sequenceable kind...
Yes, that's what I thought when you posted; just #any: But #take: is a good name
|
Re: The Inbox: Collections-mt.832.mcz
|
|
So, there are two questions: 1) Name it #take: or #any: or #slice:? 2) Signal an Error (or Notification) if there are not enough elements in the collection or just compensate (less elements or duplicate elements)? Best, Marcel
|
|
|
On Fri, May 10, 2019 at 5:53 AM Marcel Taeumel <[hidden email]> wrote:
I think the point of the original commit was to NOT signal an Error if there weren't enough elements - otherwise #first: would have worked (mostly). The idea is "I want 4, or whatever you have if you don't have that many". -cbc
|
Re: The Inbox: Collections-mt.832.mcz
|
|
> On 10.05.2019, at 17:46, Chris Cunningham <[hidden email]> wrote: > > > > On Fri, May 10, 2019 at 5:53 AM Marcel Taeumel <[hidden email]> wrote: > So, there are two questions: > > 1) Name it #take: or #any: or #slice:? > 2) Signal an Error (or Notification) if there are not enough elements in the collection or just compensate (less elements or duplicate elements)? > > I think the point of the original commit was to NOT signal an Error if there weren't enough elements - otherwise #first: would have worked (mostly). The idea is "I want 4, or whatever you have if you don't have that many". but #first does not work for Sets, Dictionaries and the like. I'm in for an Error-throwing #any:. Reason: what we have as #first on Sequenceable is more generally available on Collection as #anyOne . what we have as #first: on Sequenceable could be more generally available on Collection as #any:. Best regards -Tobias |
«
Return to Squeak - Dev
|
1 view|%1 views
| Free forum by Nabble | Edit this page |