LÖVE

I've been using the length operator (#) lots these days, since the thread here which showed it was much faster than table.maxn.

However, I've begun using structures which have both tightly-packed enumerated data, and data with string keys. For example:

test = {
x = 0,
y = 0,
name = "Funny Table",
}

for i=1, 100 do
test = i
end

this table would then have 103 pieces of data.

I often want to iterate through the numericly indexed elements of the table, which I could do with i=1, table.maxn(test). The function table.maxn is very clearly "The highest order numerical index" while the description of the length operator is a bit less clear.

Both appear to be working, but the length operator has several warnings about calling it. I can't quite make up my mind whether or not I should be safe and sacrifice some performance now. Anyone have any insight?

--Mr. Strange

As long as you don't have "holes" (i.e. x[1], x[2], x[3] and x[5] are defined, but x[4] is nil), the length operator should do just fine. My suggestion is not to have holes (in fact, my suggestion would be not to mix numbers and strings in the same table: the table's name should give you a clue about its function, and having one table with two functions means that name will be confusing whichever way you look at it).

If i remember correctly the length operator don't count the length but returns a stored length value. If you modify a list with table.insert & co it gets updated. So the length operator should work as long as you don not manually set entries like l["lala"] = 1 or l[40] = 1.
But you can always iterate over all elements and count them like "for k,v in pairs(l) do i = i +1 end" (dont know if maxn does this) but its slower than the length operator.

hagish wrote:If i remember correctly the length operator don't count the length but returns a stored length value. If you modify a list with table.insert & co it gets updated. So the length operator should work as long as you don not manually set entries like l["lala"] = 1 or l[40] = 1.
But you can always iterate over all elements and count them like "for k,v in pairs(l) do i = i +1 end" (dont know if maxn does this) but its slower than the length operator.

maxn does indeed iterate in just that fashion, which makes it appropriate but slow.

You've convinced me that I'm safe using the length operator. Thanks.

--Mr. Strange