Altered colours from subsequent use of a custom palette

[sylvaticus]

Hello, when trying to plot multiple groups of series from a dataframe and reuse the same colour palette for the individual series, the subsequent series use an "altered" version of the colours:

using DataFrames, Plots, StatPlots
x = [110,120,130,140, 105,115,125,135]
df = DataFrame(
  fruit   = ["orange","orange","orange","orange","banana","banana","banana","banana"],
  year    = [2010,2011,2012,2013,2010,2011,2012,2013],
  prod    = x,
  cons    = x - 20,
  exports = x - 40,
  imports  = x - 60,
)
plotlyjs() # same with pyplot() backend..
mycolors = [:red, :green  ] # same with mycolors = [:yellow, :orange ]
fruits_plot = plot(df, :year, :prod, group=:fruit, linestyle = :solid, linewidth=3, linecolor=:match, color_palette = mycolors)
fruits_plot = plot!(df, :year, :cons, group=:fruit, linestyle = :dot, linewidth=3, linecolor=:match, color_palette = mycolors)
fruits_plot = plot!(df, :year, :exports, group=:fruit, linestyle = :dash, linewidth=3, linecolor=:match, color_palette = mycolors)
fruits_plot = plot!(df, :year, :imports, group=:fruit, linestyle = :dashdot, linewidth=3, linecolor=:match, color_palette = mycolors)

Results in:

As you can see, the problem is that the subsequent groups of series use the palette with somehow "distorted" (or sometimes inverted) colours.

This is independent from the specific backend used.

[tbreloff]

If you want specific colors, just set the color directly (instead of match, which is the default anyways). Palette determines the "family" of colors.

…

On Thu, Feb 23, 2017 at 5:38 AM Antonello Lobianco ***@***.***> wrote: Hello, when trying to plot multiple groups of series from a dataframe and reuse the same colour palette for the individual series, the subsequent series use an "altered" version of the colours: using DataFrames, Plots, StatPlots x = [110,120,130,140, 105,115,125,135] df = DataFrame( fruit = ["orange","orange","orange","orange","banana","banana","banana","banana"], year = [2010,2011,2012,2013,2010,2011,2012,2013], prod = x, cons = x - 20, exports = x - 40, imports = x - 60, ) plotlyjs() # same with pyplot() backend.. mycolors = [:red, :green ] # same with mycolors = [:yellow, :orange ] fruits_plot = plot(df, :year, :prod, group=:fruit, linestyle = :solid, linewidth=3, linecolor=:match, color_palette = mycolors) fruits_plot = plot!(df, :year, :cons, group=:fruit, linestyle = :dot, linewidth=3, linecolor=:match, color_palette = mycolors) fruits_plot = plot!(df, :year, :exports, group=:fruit, linestyle = :dash, linewidth=3, linecolor=:match, color_palette = mycolors) fruits_plot = plot!(df, :year, :imports, group=:fruit, linestyle = :dashdot, linewidth=3, linecolor=:match, color_palette = mycolors) Results in: [image: fruits_plot] <https://cloud.githubusercontent.com/assets/2077159/23255513/3254220e-f9bc-11e6-9c95-c61933b7c8b4.png> As you can see, the problem is that the subsequent groups of series use the palette with somehow "distorted" (or sometimes inverted) colours. This is independent from the specific backend used. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#701>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA492tA9Gr29uF0-VOa45YiuEazhh6iEks5rfWGYgaJpZM4MJzao> .

[sylvaticus]

In the documentation it is stated that:

color_palette determines the colors assigned when seriescolor == :auto:

If passed a vector of colors, it will force cycling of those colors

This is what I want, as I am using it for a group of series, not for an individual one (and I am using the keywork group)
In fact, it works well on the first usage. The problem is when I add to the same plot other group of series.

Bdw, I did try to set linecolor directly, but it accepts only a scalar, not arrays..

[tbreloff]

Use a 1x2 matrix when setting colors for 2 series

…

On Thu, Feb 23, 2017 at 7:02 AM Antonello Lobianco ***@***.***> wrote: In the documentation <https://juliaplots.github.io/colors/> it is stated that: color_palette determines the colors assigned when seriescolor == :auto: If passed a vector of colors, it will force cycling of those colors This is what I want, as I am using it for a *group* of series, not for an individual one (an I am using the keywork group) In fact, it works well on the first usage. The problem is when I add to the same plot other group of series. Bdw, I did try to set linecolor directly, but it accepts only a scalar, not arrays.. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#701 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA492ldV0hYe4ipYIylorKsJb7gLRhuwks5rfXVagaJpZM4MJzao> .

[sylvaticus]

yep, using fruits_plot = plot(df, :year, :prod, group=:fruit, linestyle = :solid, linewidth=3, linecolor=[[:red] [:yellow]]) works, thanks..