How to fix “Error in `$<-.data.frame`(`*tmp*`, ”x“, value = 1L) : replacement has 1 row, data has 0” when using upsetR?

Question

Below a simulated data which will be an input to upset() function

nobs <- 4000
pb <- round(runif(n=1, min=0.1, max=0.8),1)
August <- rbinom(n=nobs, size=1, prob=pb)
pb <- round(runif(n=1, min=0.1, max=0.8),1)
September <- rbinom(n=nobs, size=1, prob=pb)
pb <- round(runif(n=1, min=0.1, max=0.8),1)
October <- rbinom(n=nobs, size=1, prob=pb)
pb <- round(runif(n=1, min=0.1, max=0.8),1)
November <- rbinom(n=nobs, size=1, prob=pb)
pb <- round(runif(n=1, min=0.1, max=0.8),1)
December <- rbinom(n=nobs, size=1, prob=pb)

#generate random id
myFun <- function(n = nobs) {
  a <- do.call(paste0, replicate(5, sample(LETTERS, n, TRUE), FALSE))
  paste0(a, sprintf("%04d", sample(9999, n, TRUE)), sample(LETTERS, n, TRUE))
}

AumPre <- round(runif(n=nobs, min=100000, max=250000))
AumAft <- round(runif(n=nobs, min=100000, max=250000))
AumIncrease <- as.integer(AumAft > AumPre)
ncust <- myFun()
df <- data.frame(ncust
             , August
             , September
             , October
             , November
             , December
             , AumPre
             , AumAft
             , AumIncrease)
upset(df, boxplot.summary = c("AumAft"))

I keep getting the following error when calling upset().

Error in `$<-.data.frame`(`*tmp*`, "x", value = 1L) : 
  replacement has 1 row, data has 0

It seems others have encountered similar issues as reported here https://github.com/hms-dbmi/UpSetR/issues/93

Hope to get a workaround on this. I like the visualisation, but couldn't make it happen when using my own datasets.

Thanks

Answer 1

I think the package may look for the first and last binary columns and consider all columns in between to be part of the sets. When you rearrange your columns, the issue goes away.

library(tidyverse)

df <- df %>% 
  select(ncust, AumPre, AumAft, August, September, October, November, December, AumIncrease)

upset(df)
upset(df, boxplot.summary = c("AumAft"))