Fix keytype and zero #574

[dpinol]

Fixes https://github.com/JuliaSparse/SparseArrays.jl/issues/574

[fredrikekre]

keytype isn't the type of the stored indices though, it should match eltype(keys(...)):

julia> v = spzeros(Int, Int32, 10)
10-element SparseVector{Int64, Int32} with 0 stored entries

julia> eltype(keys(v))
Int64

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 84.00%. Comparing base (485fd4b) to head (8c6a370).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #575      +/-   ##
==========================================
- Coverage   84.14%   84.00%   -0.15%     
==========================================
  Files          12       12              
  Lines        9160     9161       +1     
==========================================
- Hits         7708     7696      -12     
- Misses       1452     1465      +13     

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[dpinol]

keytype isn't the type of the stored indices though, it should match eltype(keys(...)):

Should I also adapt keys then, for consistency with eachindex & length?

help?> keys
search: keys keytype KeyError haskey getkey UndefKeywordError WeakKeyDict

  keys(a::AbstractArray)
Return an efficient array describing all valid indices....

[fredrikekre]

No, I think keys is correct, and thus think that keytype should always be Int.

[dpinol]

No, I think keys is correct, and thus think that keytype should always be Int.

why?

[ViralBShah]

What should do further here?

[ViralBShah]

@fredrikekre Should this be closed?