extra radii in some pie plots

[leonardslog]

Hello,

Thank you for the incredibly helpful package. I am encountering an issue while trying to display spatial admixture patterns in a dataset of 34 samples representing 3 populations. a few of which have the same coordinates, which shouldn't be a problem. 4 of the samples seem to have extra radii, splitting what should be a continuous slice of the same color into multiple ones. All rows of admixture proportions seem to appropriately sum to 1. Changing the alpha value reveals that some of these might be the samples that have additional samples in the same location, but that doesn't explain the extra slices, since with no transparency, the samples simply shouldn't be visible. The columns representing the admixture proportions are columns 2:4 (X3, X1, X2), while the long and lat coordinates are columns 5:6 (V2, V3).

Any idea what's happening? Code is below and I've attached the data and the resulting plot.

EDIT: just saw that this is addressed in #36 , so feel free to delete. I will add that upon playing with the data, I found this occurs in samples that have share only one of the same coordinates, not both, and rounding (in my case to the 100th decimal place) seems to help, though might be impractical in larger datasets where spatially overlapping samples might have meaningful and unique admixture patterns.

library(ggplot2)
library(scatterpie)
library(ggnewscale)

df <- read.csv("df.csv")

ggplot() +
  new_scale_fill() +
  geom_scatterpie(aes(x=V2, y=V3, r = 0.25),
                  data = df, cols = colnames(df[,2:4])) +
  theme_classic() +
  theme(legend.position = "none",
        axis.line = element_line(linewidth = 0, colour = NULL),
        panel.border = element_rect(linewidth = 1, linetype = "solid", colour = "black", fill = NA))

df.csv