3
\$\begingroup\$

I created various functions for string manipulation in Lua, where the strings are composed of tokens separated by an ASCII character (which may be the point, comma, semicolon etc.). One of these is the function puttok()

local function tokenize(C, strng)
    local sInput = strng or ""
    local sChar = string.format('%c', C)
    local tReturn = {}
    for sWord in string.gmatch(sInput, "[^"..sChar.."]+") do
        table.insert(tReturn, tonumber(sWord) or sWord)
    end
    return tReturn
end

local function puttok(text,token,N,C)
    local char = string.format("%c", C)
    local n
    local result
    local tokens = tokenize(C, text)
    if (N == 0) or (N > #tokens) then 
        result = text
    end
    if (N) then
        n = (N > 0) and N or #tokens + N + 1
        table.remove(tokens,n)
        table.insert(tokens,n,token)
        result = table.concat(tokens,char)
    end
    return result
end

Where

  • text = string to manipulate
  • token = token (string) we want insert
  • N = position in which will be added to the token
  • C = ASCII code of the token separator

This function is to replace a token with another. Here's an example

local text = "Violets.are.white"
text = puttok(text,"blue",3,46)
print(text)
—› Violets.are.blue

The token "blue" was inserted in position 3 in place of "white," and "Violets" and "are" are respectively in position 1 and 2. The tokens are separated by a dot (.) with ASCII code 46.

Is there a way to improve this code?
Would metatables be usefull in this case?

\$\endgroup\$
2
\$\begingroup\$
  1. Don't use the expensive string.format where string.char can do it all.

  2. You should accept a string (of length 1) as separator in tokenize and puttok, and have a default.

  3. Swapping the middle two arguments of puttok allows you to default the replacement to "".

  4. Your second conditional is useless, as it will always be taken: The first one results in a runtime-error if N is not a number.
    I wonder why you don't use an early return there anyway?

  5. First removing an element and then inserting a different one at the same index is really inefficient, resulting in a need to move all higher-numbered elements first down and then back.
    Just replace it already:

    tokens[n] = token
    
  6. Try for better names, so you don't need to explain so much in comments:

    tokenize(C, strng) => explode(delimiter, text)
    puttok(text,token,N,C) => replace_token(text, which, replacement, delimiter)
    -- Switched the middle two parameters of puttok/replace_token
    
\$\endgroup\$
  • \$\begingroup\$ Would metatables be usefull in this case? \$\endgroup\$ – user3204810 Nov 4 '15 at 20:40
  • \$\begingroup\$ I don't see what you would want to use them for, so tentatively no. \$\endgroup\$ – Deduplicator Nov 4 '15 at 20:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.