Rounding of correlation matrix causes hclust ordering differences in the new version.

[buddha2490]

I recently updated the ggcorrplot package to 0.1.2. Running ggcorrplot() on a correlation matrix with the hc.order=T option, the ordering of the heatmap drastically changed compared to the older version. I don't know if this is a bug or intentional, but I thought I should point it out to you. Between the old vs new version, it looks like the only thing updated that could affect this is that you are rounding your correlation matrix. At least in my data, this is making quite a difference in the final results.

[MichelineCampbell]

I am having the same problem with version 0.1.4. With hc.order = F, a correlation pair of r = -0.27 and p = 0.0549 is excluded (as it should be), but with hc,order = T, this correlation pair is included.