You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It's not clear whether the AbstractArray interface allows implementations to define getindex(A::AbstractArray, inds::AbstractArray) so that it returns a view of A that is affected by later mutation of A or inds. This lead to a bug in DataFrames as filter assumes that inds can be emptied, which had the unintended effect of emptying the view (JuliaData/DataFrames.jl#3192). This will be fixed soon in DataFrames anyway by copying inds, but it would be good to clarify what is expected by the interface and adjust either the manual or the filter implementation.
The text was updated successfully, but these errors were encountered:
It's not clear whether the
AbstractArray
interface allows implementations to definegetindex(A::AbstractArray, inds::AbstractArray)
so that it returns a view ofA
that is affected by later mutation ofA
orinds
. This lead to a bug in DataFrames asfilter
assumes thatinds
can be emptied, which had the unintended effect of emptying the view (JuliaData/DataFrames.jl#3192). This will be fixed soon in DataFrames anyway by copyinginds
, but it would be good to clarify what is expected by the interface and adjust either the manual or thefilter
implementation.The text was updated successfully, but these errors were encountered: