Skip to content

Avoid Union{Missing,T} in columns after DropMissing #35

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Apr 9, 2022
Merged

Avoid Union{Missing,T} in columns after DropMissing #35

Show file tree
Hide file tree
Changes from 5 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
6daefd6
Avoid Union{Missing,T}
eliascarv Apr 4, 2022
f926ec8
Add _process function
eliascarv Apr 4, 2022
a041ca6
Add _reverttypes
eliascarv Apr 5, 2022
4826e04
Avoid unnecessary calls to the collect function in _nonmissing
eliascarv Apr 5, 2022
3aeea96
Add tests
eliascarv Apr 5, 2022
0c92b16
Move post and pre processing to apply and revert
eliascarv Apr 5, 2022
59f32ca
Use Tables.columntable to simplify the code.
eliascarv Apr 6, 2022
7db5e28
Apply suggestions p1
eliascarv Apr 7, 2022
1514725
Apply suggestions p2
eliascarv Apr 8, 2022
3634072
Apply suggestions p3
eliascarv Apr 9, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
53 changes: 41 additions & 12 deletions src/transforms/filter.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,10 +36,6 @@ function revert(::Filter, newtable, cache)
rows |> Tables.materializer(newtable)
end

# DropMissing

const VecOrTuple{T} = Union{Vector{T}, NTuple{N, T}} where {T, N}

"""
DropMissing()
DropMissing(:)
Expand Down Expand Up @@ -69,22 +65,55 @@ DropMissing(cols::T...) where {T<:ColSelector} =

isrevertible(::Type{<:DropMissing}) = true

_ftrans(::DropMissing{Colon}, table) =
# ftrans
_ftrans(::DropMissing{Colon}, cols) =
Filter(row -> all(!ismissing, row))

function _ftrans(transform::DropMissing, table)
allcols = Tables.columnnames(table)
cols = _filter(transform.colspec, allcols)
_ftrans(::DropMissing, cols) =
Filter(row -> all(!ismissing, getindex.(Ref(row), cols)))

# nonmissing
_nonmissing(::Type{T}, c) where {T} = c
_nonmissing(::Type{Union{Missing,T}}, c) where {T} =
collect(T, c)

function _nonmissing(columns, col)
c = Tables.getcolumn(columns, col)
_nonmissing(eltype(c), c)
end

function _nonmissing(table, cols, allcols)
columns = Tables.columns(table)
newcols = [col ∈ cols ? _nonmissing(columns, col) : Tables.getcolumn(columns, col)
for col in allcols]
𝒯 = (; zip(allcols, newcols)...)
𝒯 |> Tables.materializer(table)
end

# reverttypes
function _reverttypes(table, types)
columns = Tables.columns(table)
allcols = Tables.columnnames(table)
newcols = [collect(T, Tables.getcolumn(columns, col))
for (T, col) in zip(types, allcols)]
𝒯 = (; zip(allcols, newcols)...)
𝒯 |> Tables.materializer(table)
end

function apply(transform::DropMissing, table)
ftrans = _ftrans(transform, table)
allcols = Tables.columnnames(table)
cols = _filter(transform.colspec, allcols)
ftrans = _ftrans(transform, cols)
newtable, fcache = apply(ftrans, table)
newtable, (ftrans, fcache)
# post-processing
types = Tables.schema(table).types
ptable = _nonmissing(newtable, cols, allcols)
ptable, (ftrans, fcache, types)
end

function revert(::DropMissing, newtable, cache)
ftrans, fcache = cache
revert(ftrans, newtable, fcache)
ftrans, fcache, types = cache
# pre-processing
ptable = _reverttypes(newtable, types)
revert(ftrans, ptable, fcache)
end
39 changes: 39 additions & 0 deletions test/transforms.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -598,6 +598,45 @@
@test isequalmissing(n.e, [missing, 5, 6, 5])
@test isequalmissing(n.f, [4, missing, 4, 5])

# column eltype
ttypes = Tables.schema(t).types

T = DropMissing()
n, c = apply(T, t)
ntypes = Tables.schema(n).types
@test ntypes[1] == Int
@test ntypes[2] == Int
@test ntypes[3] == Int
@test ntypes[4] == Int
@test ntypes[5] == Int
@test ntypes[6] == Int
tₒ = revert(T, n, c)
@test ttypes == Tables.schema(tₒ).types

T = DropMissing([:a, :c, :d])
n, c = apply(T, t)
ntypes = Tables.schema(n).types
@test ntypes[1] == Int
@test ntypes[2] == Union{Missing,Int}
@test ntypes[3] == Int
@test ntypes[4] == Int
@test ntypes[5] == Union{Missing,Int}
@test ntypes[6] == Union{Missing,Int}
tₒ = revert(T, n, c)
@test ttypes == Tables.schema(tₒ).types

T = DropMissing([:b, :e, :f])
n, c = apply(T, t)
ntypes = Tables.schema(n).types
@test ntypes[1] == Union{Missing,Int}
@test ntypes[2] == Int
@test ntypes[3] == Union{Missing,Int}
@test ntypes[4] == Union{Missing,Int}
@test ntypes[5] == Int
@test ntypes[6] == Int
tₒ = revert(T, n, c)
@test ttypes == Tables.schema(tₒ).types

# reapply test
T = DropMissing()
n1, c1 = apply(T, t)
Expand Down