Bump compat for CategoricalArrays to 0.8

NEWS.md

@@ -3,6 +3,7 @@ Each release typically has a number of minor bug fixes beyond what is listed her
 
 # Version 1.x
 
+ * Add support for CategoricalArrays v0.8 & and DataFrames v0.21 (#1447)
  * Add `Guide.manual_discrete_key`, updates `Guide.manual_color_key` (#1441)
  * Increase support for syntax `color=[colorant"color"]` (#1438)
  * Enable `color` aesthetic for `Stat.qq` (#1434)

Project.toml

@@ -31,12 +31,12 @@ Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 
 [compat]
 Cairo = "0.7, 0.8, 1.0"
-CategoricalArrays = "0.5, 0.6, 0.7"
+CategoricalArrays = "0.8"
 Colors = "0.9, 0.10, 0.11, 0.12"
 Compose = "0.7, 0.8"
 Contour = "0.5"
 CoupledFields = "0.1, 0.2"
-DataFrames = "0.15, 0.16, 0.17, 0.18, 0.19, 0.20"
+DataFrames = "0.21"
 DataStructures = "0.11, 0.12, 0.13, 0.14, 0.15, 0.16, 0.17"
 Distributions = "0.16, 0.17, 0.18, 0.19, 0.20, 0.21, 0.22, 0.23"
 DocStringExtensions = "0.7, 0.8"

docs/Project.toml

@@ -12,5 +12,5 @@ RDatasets = "ce6b1742-4840-55fa-b093-852dadbb1d8b"
 Showoff = "992d4aef-0814-514b-bc4d-f2e9a6c4116f"
 
 [compat]
-CSV = "0.4, 0.5"
+CSV = "0.6"
 Documenter = "~0.22"

docs/src/gallery/geometries.md

@@ -60,7 +60,8 @@ hstack(p1, p2, p3)
 using Gadfly, RDatasets, DataFrames
 set_default_plot_size(21cm, 8cm)
 
-D = by(dataset("datasets","HairEyeColor"), [:Eye,:Sex], Frequency=:Freq=>sum)
+hecolor = dataset("datasets","HairEyeColor")
+D = combine(groupby(hecolor, [:Eye,:Sex]), :Freq=>sum=>:Frequency)
 p1 = plot(D, color=:Eye, y=:Frequency, x=:Sex, Geom.bar(position=:dodge))
 
 palette = ["brown","blue","tan","green"]  # Is there a hazel color?
@@ -235,19 +236,19 @@ set_default_plot_size(21cm, 8cm)
 salaries = dataset("car","Salaries")
 salaries.Salary /= 1000.0
 salaries.Discipline = ["Discipline $(x)" for x in salaries.Discipline]
-df = by(salaries, [:Rank,:Discipline], :Salary=>mean, :Salary=>std)
-df.ymin, df.ymax = df.Salary_mean.-df.Salary_std, df.Salary_mean.+df.Salary_std
-df.label = string.(round.(Int, df.Salary_mean))
+fn1(x, u=mean(x), s=std(x)) = (Salary=u, ymin=u-s, ymax=u+s, 
+    label="$(round.(Int,u))")
+df = combine(:Salary=>fn1, groupby(salaries, [:Rank, :Discipline]))
 
-p1 = plot(df, x=:Discipline, y=:Salary_mean, color=:Rank,
+p1 = plot(df, x=:Discipline, y=:Salary, color=:Rank,
     Scale.x_discrete(levels=["Discipline A", "Discipline B"]),
     ymin=:ymin, ymax=:ymax, Geom.errorbar, Stat.dodge,
     Geom.bar(position=:dodge),
     Scale.color_discrete(levels=["Prof", "AssocProf", "AsstProf"]),
     Guide.colorkey(title="", pos=[0.76w, -0.38h]),
     Theme(bar_spacing=0mm, stroke_color=c->"black")
 )
-p2 = plot(df, y=:Discipline, x=:Salary_mean, color=:Rank,
+p2 = plot(df, y=:Discipline, x=:Salary, color=:Rank,
     Coord.cartesian(yflip=true), Scale.y_discrete,
     xmin=:ymin, xmax=:ymax, Geom.errorbar, Stat.dodge(axis=:y),
     Geom.bar(position=:dodge, orientation=:horizontal),
@@ -630,6 +631,6 @@ hstack(p1,p2)
 using Gadfly, RDatasets
 set_default_plot_size(14cm, 8cm)
 Dsing = dataset("lattice","singer")
-Dsing.Voice = [x[1:5] for x in Dsing.VoicePart]
+Dsing.Voice = [x[1:5] for x in Array(Dsing.VoicePart)]
 plot(Dsing, x=:VoicePart, y=:Height, color=:Voice, Geom.violin)
 ```

docs/src/gallery/guides.md

@@ -48,12 +48,12 @@ p1 = plot(layer(points, x=:index, y=:val, color=[colorant"green"]),
     Guide.manual_color_key("Legend", ["Points", "Line"], ["green", "deepskyblue"],
         shape=[Shape.circle, Shape.hline]))
 
-D = dataset("COUNT", "titanicgrp")
-D = join(D, by(D, :Class, :Cases=>sum), on=:Class)
-D.prcnt = 100*D.Survive./D.Cases_sum
+D = groupby(dataset("COUNT", "titanicgrp"), :Class)
+fn1(s,c) = 100*s./sum(c)
+D = combine(D, :Age, :Sex, [:Survive, :Cases]=>fn1=>:prcnt)
 p2 = plot(stack(D, [:Age, :Sex]), xgroup=:Class,
     Geom.subplot_grid(layer(x=:variable, y=:prcnt, color=:value, Geom.bar)),
-    Scale.x_discrete, Guide.ylabel("Survival (%)"),
+    Scale.x_discrete, Guide.ylabel("Survival (% Class)"),
     Guide.manual_color_key("Age", ["children","adults"], 1:2),
     Guide.manual_color_key("Sex", ["female","male"], 3:4),
     Theme(bar_spacing=1mm, key_position=:none,
@@ -84,10 +84,9 @@ plot(Dsleep, x=:BodyWt, y=:BrainWt, Geom.point, color=:Vore, shape=:SleepTime,
 using Compose, Gadfly, RDatasets
 set_default_plot_size(14cm, 8cm)
 
-Titanic = dataset("datasets", "Titanic")
-Class =  by(Titanic, :Class, :Freq=>sum)
-Titanic = join(Titanic[Titanic.Survived.=="Yes",:], Class, on=:Class)
-Titanic.prcnt = 100*Titanic.Freq./Titanic.Freq_sum
+D = groupby(dataset("datasets", "Titanic"), :Class)
+Titanic = combine(D, :, :Freq=>(c->100*c./sum(c))=>:prcnt)
+filter!(:Survived=>x->x=="Yes", Titanic)
 sizemap = n->range(3pt, 8pt, length=n)
 
 plot(Titanic, Scale.x_log10,  Scale.y_log10,

docs/src/gallery/scales.md

@@ -119,7 +119,8 @@ plot(x=rand(12), y=rand(12), color=repeat(["a","b","c"], outer=[4]),
 ```@example
 using Gadfly, RDatasets, DataFrames
 set_default_plot_size(14cm, 8cm)
-D = by(dataset("datasets","HairEyeColor"), [:Eye,:Sex], Frequency=:Freq=>sum)
+hecolor = dataset("datasets","HairEyeColor")
+D = combine(groupby(hecolor, [:Eye,:Sex]), :Freq=>sum=>:Frequency)
 palette = ["brown","blue","tan","green"] # Is there a hazel color?
 pa = plot(D, x=:Sex, y=:Frequency, color=:Eye, Geom.bar(position=:stack),
           Scale.color_discrete_manual(palette...))
@@ -169,9 +170,9 @@ using Gadfly, RDatasets, Statistics
 set_default_plot_size(14cm, 8cm)
 
 tips = dataset("reshape2", "tips")
-tipsm = by(tips, [:Day, :Sex], :TotalBill=>mean, :Tip=>mean)
+tipsm = combine(groupby(tips, [:Day, :Sex]), [:TotalBill, :Tip].=>mean)
 
- plot(tipsm, Geom.point,
+plot(tipsm, Geom.point,
     x=:TotalBill_mean, y=:Tip_mean, color=:Sex, shape=:Day,
     layer(x=:TotalBill_mean, y=:Tip_mean, group=:Day, Geom.line,
         style(default_color=colorant"gray")),
@@ -210,10 +211,9 @@ plot(aq, x=:Day, y=:Ozone, color=:Month, size=:Wind,
 using Compose, Gadfly, RDatasets
 set_default_plot_size(14cm, 8cm)
 
-Titanic = dataset("datasets", "Titanic")
-Class =  by(Titanic, :Class, :Freq=>sum)
-Titanic = join(Titanic[Titanic.Survived.=="Yes",:], Class, on=:Class)
-Titanic.prcnt = 100*Titanic.Freq./Titanic.Freq_sum
+D = groupby(dataset("datasets", "Titanic"), :Class)
+Titanic = combine(D, :, :Freq=>(c->100*c./sum(c))=>:prcnt)
+filter!(:Survived=>x->x=="Yes", Titanic)
 sizemap = n->range(4pt, 12pt, length=n)
 
 plot(Titanic, Scale.x_log10,  Scale.y_log10,
@@ -255,7 +255,7 @@ Random.seed!(1234)
 p1 = plot(x=rand(1:3, 20), y=rand(20), Scale.x_discrete)
 # To perserve the order of the columns in the plot when plotting a DataFrame
 df = DataFrame(v1 = randn(10), v2 = randn(10), v3 = randn(10))
-p2 = plot(df, x=Col.index, y=Col.value, Scale.x_discrete(levels=names(df)))
+p2 = plot(df, x=Col.index, y=Col.value, Scale.x_discrete(levels=propertynames(df)))
 hstack(p1,p2)
 ```
 

docs/src/gallery/statistics.md

@@ -41,8 +41,7 @@ set_default_plot_size(21cm, 8cm)
 salaries = dataset("car","Salaries")
 salaries.Salary /= 1000.0
 salaries.Discipline = ["Discipline $(x)" for x in salaries.Discipline]
-df = by(salaries, [:Rank,:Discipline], :Salary=>mean, :Salary=>std)
-df.ymin, df.ymax = df.Salary_mean.-df.Salary_std, df.Salary_mean.+df.Salary_std
+df = combine(groupby(salaries, [:Rank, :Discipline]), :Salary.=>mean)
 df.label = string.(round.(Int, df.Salary_mean))
 
 p1 = plot(df, x=:Discipline, y=:Salary_mean, color=:Rank, 
@@ -87,7 +86,7 @@ plot(y=[sigmoid, x->sigmoid(x+2)], xmin=[-10], xmax=[10],
 using Distributions, Gadfly, RDatasets
 set_default_plot_size(21cm, 8cm)
 iris, geyser = dataset.("datasets", ["iris", "faithful"])
-df = by(iris, :Species, d=:SepalLength=>x->fit(Normal, x))
+df = combine(groupby(iris, :Species), :SepalLength=>(x->fit(Normal, x))=>:d)
 ds2 = fit.([Normal, Uniform], [geyser.Eruptions])
 
 yeqx(x=4:6) = layer(x=x, Geom.abline(color="gray80"))

src/dataframes.jl

@@ -4,7 +4,7 @@ function meltdata(U::AbstractDataFrame, colgroups::Vector{Col.GroupedColumn})
     um, un = size(U)
 
     # Figure out the size of the new melted matrix
-    allcolumns = Set{Symbol}(names(U))
+    allcolumns = Set{Symbol}(propertynames(U))
 
     vm = um
     colidxs = [colgroup.columns===nothing ? collect(allcolumns) : colgroup.columns for colgroup in colgroups]
@@ -18,7 +18,7 @@ function meltdata(U::AbstractDataFrame, colgroups::Vector{Col.GroupedColumn})
     vnames = Symbol[]
     colmap = Dict{Any, Int}()
 
-    eltypd = Dict(k=>v for (k,v) in zip(names(U), eltype.(eachcol(U))))
+    eltypd = Dict{Symbol, DataType}(k=>v for (k,v) in zip(propertynames(U), eltype.(eachcol(U))))
     # allocate vectors for grouped columns
     for (j, (colgroup, colidx)) in enumerate(zip(colgroups, colidxs))
         eltyp = promote_type(getindex.([eltypd], colidx)...)

src/misc.jl

@@ -331,7 +331,7 @@ discretize_make_ia(values::CategoricalArray) =
     discretize_make_ia(values, intersect(push!(levels(values), missing), unique(values)))
 discretize_make_ia(values::CategoricalArray, ::Nothing) = discretize_make_ia(values)
 function discretize_make_ia(values::CategoricalArray{T}, levels::Vector) where {T}
-    mapping = something.(indexin(CategoricalArrays.index(values.pool), levels), 0)
+    mapping = something.(indexin(CategoricalArrays.levels(values.pool), levels), 0)
     pushfirst!(mapping, something(findfirst(ismissing, levels), 0))
     index = [mapping[x+1] for x in values.refs]
     any(iszero, index) && throw(ArgumentError("values not in levels encountered"))

test/testscripts/issue975.jl

@@ -1,6 +1,6 @@
 using DataFrames, Gadfly, Dates
 
-set_default_plot_size(6inch, 3inch)
+set_default_plot_size(7inch, 3inch)
 
 # Geom.smooth already has tests for these types:
 # Floats: smooth_lm.jl
@@ -15,16 +15,16 @@ n = length(t)
 D = DataFrame(t1=t1, t2=t2,
               cycle = 5*sin.(t*2π/365.25)+randn(n),
               trend = 0.1*[1.0:n;].+2*randn(n))
-Dl = stack(D, setdiff(names(D), [:t1,:t2]))
+Dl = stack(D, Not([:t1, :t2]))
 
-plot(D,
-    x=:t1, y=:trend, Geom.point,
-    Geom.smooth(method=:lm),
+p1 = plot(D, x=:t1, y=:trend, 
+    Geom.smooth(method=:lm), Geom.point,
     Theme(point_size=1.8pt, key_position=:none)
     )
 
-plot(Dl,
-    x=:t2, y=:value, color=:variable, Geom.point,
-    Geom.smooth(smoothing=0.05),
+p2 = plot(Dl,
+    x=:t2, y=:value, color=:variable, 
+    Geom.smooth(smoothing=0.05), Geom.point,
     Theme(point_size=1.8pt, key_position=:none)
     )
+hstack(p1, p2)

test/testscripts/stat_dodge.jl

@@ -4,7 +4,7 @@ set_default_plot_size(21cm, 8cm)
 salaries = dataset("car","Salaries")
 salaries.Salary /= 1000.0
 salaries.Discipline = ["Discipline $(x)" for x in salaries.Discipline]
-df = by(salaries, [:Rank,:Discipline], :Salary=>mean, :Salary=>std)
+df = combine(groupby(salaries, [:Rank,:Discipline]), :Salary=>mean, :Salary=>std)
 df.ymin, df.ymax = df.Salary_mean.-df.Salary_std, df.Salary_mean.+df.Salary_std
 df.label = string.(round.(Int, df.Salary_mean))
 

test/testscripts/wide_form.jl

@@ -4,5 +4,5 @@ set_default_plot_size(6inch, 4inch)
 
 births = dataset("HistData", "Arbuthnot")
 
-plot(births, x=Col.value(:Year), y=Col.value(:Males, :Females),
+plot(births, x=:Year, y=Col.value(:Males, :Females),
      color=Col.index(:Males, :Females), Geom.line)