Big doubt..

[Feche1320]

When and where should I use 'pairs' or 'ipairs'?

For example:

  
local table = { "hello", "i", "am", "here", "not" } 
for i, tmp in ipairs(table) do 
    outputChatBox(tmp) 
end 

  
local table = { "hello", "i", "am", "here", "not" } 
for i, tmp in pairs(table) do 
    outputChatBox(tmp) 
end 

Will print the same:

'hello

i

am

here

not'

So.. differences?

Thanks

[DakiLLa]

ipairs is used to go through a table with integer indexes; pairs to go through string indexes.

local t = { "hello", "i", "am" } 
for i, v in ipairs( t ) do 
  print( v ); 
end -- I am 
  
local t = { ["hello"] = "I am", ["hello_again"] = "not here" } 
--local t = { hello = "I am", hello_again = "not here" } is equivalent of the previous line 
for i, v in pairs( t ) do 
  print( v ); 
end -- I am not here 

By the way, you can't use pairs and ipairs to loop through tables with mixed types of indexes.

[proracer]

ipairs can be used to go trough string indexes also if I'm not wrong...

[Aibo]

ipairs will iterate the table by numeric index order starting from 1 and until next numeric key is not found, like:

table = { "one", "two", "three", four = "four", five = "five", [5] = "five", six = "six" }

ipairs(test) will stop after table[3] because table[4] is nil. and it will not find test[5], because [4] is missing.

and pairs will output everything with any kind of index, be it number, string or MTA element.

[proracer]

Also I sometimes don't understand this, I saw many people use ipairs and with your example I think pairs is the best way.

What is the reason?

[karlis]

ipairs is basically used where order is needed,

while pairs iterates tru table completely random, thus making it a good random generator too.

[Zango]

pairs versus ipairs on numerically indexed tables is significantly slower.

[Feche1320]

Thanks!