Changed Dict(ks,vs) to behave like Dict(zip(ks,vs))

[ivarne]

Currently the lenght of ks is used to deterimine the numer of elements in the dictionary. After this, the shortest of ks and vs is used.

An alternative might be to throw an error if they are not equal length, but I think the same behavior as Dict(zip(ks,vs)), is the most intuitive.

[toivoh]

I would vote to throw an error for unequal lengths both here and with zip. Or is anyone actually using that behavior of zip intentionally?

[ivarne]

The behavior with unequal lengths could be used with infinite iterators, but I have not seen any of those in Julia.

Examples might be a range that never stops counting (just wraps around) or something that just repeats a number, but I am not aware of any uses in Julia.

Dict(zip(keys, repeat(0))) would be a useful idem to create a dictionary with keys initialized to zero.

Hmm... I just realized that Dict does not have a 1 argument constructor for iterables that would be suitable for the zip idiom. Is that a oversight, or is there any reasoning behind?

[ivarne]

Oops. the last comment is discussed in #4871

[JeffBezanson]

We might as well just switch to zip, which now runs at full speed for 2 arguments. But, length can serve as a size hint (it was silly of me not to use this in the current constructor), and general iterators might not support length.