Re: pairs(t, skey) and ipairs(t, skey)

[Andrew Starks <andrew.starks@...>]

On Wed, Oct 2, 2013 at 4:29 PM, Sean Conner <sean@conman.org> wrote:
> Given:
>
>         t_seq = {}
>         for i = 1 , 99 do
>           t_seq[i] = nil
>         end
>         t_seq[100] = "Boom"
>
>         t_hole = {}
>         t_hole[100] = "Boom"
>
>   What should the following print?
>
>         print(#t_seq)
>         print(#t_hole)

Given that `nil` is used to denote a key that is "not there", the
above would not be a sequence.

Using Tim's example, I'd say that your two print statements should print "nil".

If I may channel Tim's intent for a moment:


EMPTY = {}

t_seq = {}
for i = 1, 99 do
    t_seq[i] = EMPTY
end

t_seq[100] = "BOOM"

t_empty = {}

t_hash = { foo= "baz", bar = "dizzle"}
t_hybrid = {"first", "second", "third", hash = "value }

t_broken_sequence = {"first", "second", [4] = "fourth"}
print(#t_seq)
--> 100
print(#t_empty)
--> 0
print(#t_hash)
--> 0
print(#hybrid)
--> 3
print(#broken_sequence)
--> nil

Or something like it. Again, more of that "frowned upon" type of
proposal. It's not so important that "nil"  be returned. It's more
important that the regular rules apply: nil is not a value that you
can effectively store in a table, apart from using it to know if that
index is non-existant.

-Andrew