Re: [Bug-apl] West way to remove substrings?

[enztec]

something i had from a while back 


#! rmsubstr

∇rs←substr rmsubstr string ⍝ " " quotes required for all strings and substrings

⍝ test variables - names for ease in testing
 s1←string1←"abcdefghijkcdecef"
 s2←string2←"abcdefghijkcdecefa"
 s3←string3←"a"    ⍝ single element string
 ss1←substr1←"cde" ⍝ present in string 2x returns string with substring removed
 ss2←substr2←"cee" ⍝ substring not there returns full string
 ss3←substr3←"efa" ⍝ substring there in s2/string2 but not s1/string1
 ss4←substr4←"a"   ⍝ single element substr
 ss4←substr5←""    ⍝ substring not there returns full string (special test for 
0=⍴,substring)

 r←(Rs←⍴string)-¯1+Rss←⍴substr
 
 ⍎((0=Rs)∨(0=Rss))/'''either string and/or substring are not given'' ◊ 
rs←string ◊ →0'

 ⍎(Rss>Rs)/'''substring is bigger then string'' ◊ rs←string ◊ →0'
⍝ 'r : ', r, '  Rs : ', Rs, '  Rss : ', Rss

 ssa←((r,Rss)⍴substr)
⍝ 'ssa : '
⍝ ssa

 sa←(string[((r,Rss)⍴¯1+⍳Rss)+⍉(Rss,r)⍴⍳r])
⍝ 'sa : '
⍝ sa

 'string : ', string ⍝ comment for use as working function
 'substring to remove from string : ', substr ⍝ comment for use as working 
function

 di←(Rss=+/ssa=sa)/⍳r
⍝ di←(Rss=ssa+.=sa)/⍳r ⍝ inner product fails here but should work ?

 ⍎(0=⍴di)/'rs←string ◊ →0' ⍝ substring not found - intact string returned

 o←Rs⍴1
 o[,(((⍴di), Rss)⍴¯1+⍳Rss)+⍉(Rss,⍴di)⍴di]←0

 rs←o/string

∇




On Tue, 14 Mar 2017 22:20:42 +0800
Elias Mårtenson <address@hidden> wrote:

> In a discussion on the #lisp IRC channel recently, someone wanted to remove
> a certain substring from a larger string. In Lisp, the solution is to
> simply use cl-ppcre and remove by regex.
> 
> But, I started thinking about how to do this in APL, and this is what I
> came up with:
> 
>       RemoveSubstrings ← {,/ ⍵ ⊂⍨ ↑~∨/ (-⍳⍴⍺) °.⌽ ⊂⍺⍷⍵}
> 
> Is this a good way of solving the problem? I realise that I have a bug when
> the resulting string is empty:
> 
> *      "foo" RemoveSubstrings "foo"*
> DOMAIN ERROR
> λ1[1]  λ←,/⍵⊂⍨↑∼∨/(-⍳⍴⍺)°.⌽⊂⍺⋸⍵
>          ^ ^
> 
> Does anyone have a better solution?
> 
> Regards,
> Elias