Divran's string.Explode tests

explodetest.lua

local totable = string.ToTable
function Explode1(separator, str, withpattern)
    if (separator == "") then return totable( str ) end
    
    local ret = {}
    local index,lastPosition = 1,1
    
    -- Escape all magic characters in separator
    if not withpattern then separator = separator:gsub( "[%-%^%$%(%)%%%.%[%]%*%+%-%?]", "%%%1" ) end
    
    if (#str > 10000) then
    
        -- Find the parts
        for startPosition,endPosition in str:gmatch( "()" .. separator.."()" ) do
            ret[index] = str:sub(lastPosition, startPosition-1)
            index = index + 1
            
            -- Keep track of the position
            lastPosition = endPosition
        end
        
    else
    
        -- Find the parts
        for part,endPosition in str:gmatch( "(.-)" .. separator.."()" ) do
            ret[index] = part
            index = index + 1
            
            -- Keep track of the position
            lastPosition = endPosition
        end
    
    end
    
    
    -- Add last part by using the position we stored
    ret[index] = str:sub(lastPosition)
    return ret
end

function Explode2(separator, str, withpattern)
    if (separator == "") then return totable( str ) end
    
    local ret = {}
    local index,lastPosition = 1,1
    
    -- Escape all magic characters in separator
    if not withpattern then separator = separator:gsub( "[%-%^%$%(%)%%%.%[%]%*%+%-%?]", "%%%1" ) end
    
    -- Find the parts
    for startPosition,endPosition in str:gmatch( "()" .. separator.."()" ) do
        ret[index] = str:sub(lastPosition, startPosition-1)
        index = index + 1
        
        -- Keep track of the position
        lastPosition = endPosition
    end
    
    -- Add last part by using the position we stored
    ret[index] = str:sub(lastPosition)
    return ret
end

function Explode3(separator, str, withpattern)
    if (separator == "") then return totable( str ) end
    
    local ret = {}
    local index,lastPosition = 1,1
    
    -- Escape all magic characters in separator
    if not withpattern then separator = separator:gsub( "[%-%^%$%(%)%%%.%[%]%*%+%-%?]", "%%%1" ) end
    
    -- Find the parts
    for part,endPosition in str:gmatch( "(.-)" .. separator.."()" ) do
        ret[index] = part
        index = index + 1
        
        -- Keep track of the position
        lastPosition = endPosition
    end
    
    -- Add last part by using the position we stored
    ret[index] = str:sub(lastPosition)
    return ret
end

-- I used these strings to test

-- "long" string:
local str = ("aaaaaaaaaaaaaaaaaaaaaaa"):rep(2000)
str = str:sub(1,math.floor(#str/2)) .. "b" .. str:sub(math.floor(#str/2)+1)
local sep = "b"

-- "short" string:
local str = ("aab"):rep(100) -- test 1
local str = ("aab"):rep(20000) -- test 2
local str = ("aab"):rep(40000) -- test 3

--[[ Test results:
Test 1: The "long string" which has massive parts and only one separator. The explode function is called only ONCE, and not in a loop.
-------------------------------------
long string char nr: 46001
Explode1: 0.001220703125
Explode2: 0.001220703125
Explode3: 4.3232421875
As you can see, #3 takes over 4 seconds to explode this string.

Test 2: The "short string" which consists of "aab" repeated 100 times, and the explode functions looped 10 000 times.
-------------------------------------
short string char nr: 300
Explode1: 0.4273681640625
Explode2: 0.581787109375
Explode3: 0.375

Test 3: The "short string" repeat has been increased to 20 000, and the loops have been commented out (so this is run ONCE).
-------------------------------------
short string char nr: 60000
Explode1: 0.01123046875
Explode2: 0.010498046875
Explode3: 0.0057373046875

Test 4: The "short string" repeat has been increased to 40 000, and the loops are still commented out.
-------------------------------------
short string char nr: 120000
Explode1: 0.0240478515625
Explode2: 0.022705078125
Explode3: 0.013427734375
]]