OpenEuphoria

Re: better flatten()?

• Posted by petelomax Jul 27, 2015
• 1553 views

The revised version above yields the much more reasonable "abc, def, ghi, jkl, mno, pqr, stu, vwx, yz", thanks

Without adequate testing, may I quickly suggest:

-- 
-- string type borrowed from std/types.e 
-- 
type string( object x )  -- unchanged 
     
    if not sequence(x) then 
        return 0 
    end if 
     
    for i = 1 to length(x) do 
        if not integer(x[i]) then 
            return 0 
        end if 
        if x[i] < 0 then 
            return 0 
        end if 
        if x[i] > 255 then 
            return 0 
        end if 
    end for 
     
    return 1 
end type 
 
-- 
-- flatten a sequence, preserving nested strings 
-- 
function flatten_seq2( sequence s1, string delim = "" )  
      
sequence s2 = ""  
object s1i 
      
    for i = 1 to length( s1 ) do  
        if i!=1 then  
            s2 &= delim  
        end if  
        s1i = s1[i] 
        if atom( s1i ) or string(s1i) then  -- (edit: wrong test) 
            s2 &= s1i  
              
        else  
            s2 &= flatten_seq2( s1i,delim )  
              
        end if  
          
    end for  
      
    return s2  
end function  

That string_array thing just looked wrong to me, like it was only ever going to work at a single nested level. Of course flatten() needs to do something sensible when passed a tree of words.

Oh, I had to remove join because I was making it recursive, it now calls itself rather than flatten_all, and that just feels more sensible.

I'm thinking it is reasonable to say that when processing non-string stuff you should not use a delimiter, and if you do specify one there is a fairly strong implication that you are expecting a single string back.

To that end I sneaked in a type check on delim.

I'm not absolutely against the idea of two functions, but I'm not convinced we really need that split yet either.

Pete