issue with reshape masking "melt" in ggcorrplot

[kassambara]

(e-mail from a user)

I’m using your “ggcorrplot” package, and I believe I have discovered a bug that occurs if the user also has the “reshape” package loaded.

Within your “ggcorrplot” function, you use reshape2::melt to reshape the correlation matrix into long format with this line:

corr <- reshape2::melt(corr, na.rm = TRUE)

You then plot the data by referring to variable names created by reshape2, “Var1” and “Var2”:

     p <- ggplot2::ggplot(corr, ggplot2::aes_string("Var1", "Var2", fill = "value"))

However, if the user also has the “reshape” package loaded, the variables are called “X1” and “X2” by default rather than “Var1” and “Var2.” This occurs even if “reshape2” is also loaded, due to the “melt” function being masked.

I am writing a function that requires using both the “reshape” package and your “ggcorrplot” function, and this causes a problem because I cannot run your function due to this bug. Would you be interested in investigating and correcting this problem?