No strict ticks found message

Specific Domains Visualization
1

I was plotting and realised a message was popping up in the REPL. I am using Julia 1.5 in Atom and StatsPlots.jl and Plot.jl. Any idea what is causing it?

 Warning: No strict ticks found
└ @ PlotUtils C:\Users\Helga\.julia\packages\PlotUtils\nCtbM\src\ticks.jl:283

image

2

Could be a float precision thing? what are the numbers like?

3

Precision or large numbers are definitely an issue, at least for the GR backend.
works:
plot([1,2,3],[1,1,1]*1E8)
fails:
plot([1,2,3],[1,1,1]*1E16)

4

The numbers don’t seem too strange but there are a lot of decimal places. I don’t need that much accuracy though.

0.695873142253793

Is it possible it comes about from plotting too much? The previous plot came out fine

image

5

Is it possible there is a “strict” that is in the code that should be “struct”?

6

The message appears when there is :log10 in the code

scatter(df_high_ranges_final_1[!, :clearance],diag(Matrix(df_high_ranges_final_1[:, (33):(42)])),xaxis = :log10,yaxis = :log10, xlabel = "Vessel Coverage(1*m-2)", ylabel = "Final Target Biomass (g*m-2)", label = "fw1", leg = :outerright)
7

Does your data have 0’s?

2 Likes
8

Yes, I have some 0’s in my vector…

1 Like
9

Well, I’m getting the same warning and my data definitely doesn’t have zeros since I manually shift all zeros to 1e-16. Anything else that could cause this?

10

Using older versions of Plots.jl.

11

I’m not sure. I have Plots 1.23, which is not the latest but my registry won’t update to 1.25. As simple an example as this reproduces my problem:

x = range(1, 2*pi, length=100)
P = plot(x, 1.00001 .+ cos.(x), yaxis = :log10, xaxis = :log10)

┌ Warning: No strict ticks found
└ @ PlotUtils /Users/ortner/.julia/packages/PlotUtils/VgXdq/src/ticks.jl:294
12

Plots fine here using Plots; gr()
(Plots v1.25.0)

Activate a temp enviroment and test it.

13

I get the warning with axis=:log10; I do not get the warning with xaxis=:log10, yaxis=:log10. I thought these two should be equivalent, no?

Same goes for scale, xscale and yscale.

14

I think they should be equivalent. Note that the warning only seems to concern the gr() backend. Others, such as plotly() or pythonplot(), don’t complain.