grand_summary_rows errors on all NA columns

[smingerson]

latest from github (GithubSHA1: 883df9a)

My use case is an empty table showing what information will be displayed. The end user will enter info in a Shiny app and then the table will update. My workaround will be to populate the table with 0s. Edit: this also happens with 0. The error happened in the print method, when it is applying the functions to the columns.

library(gt)
grand_summary_rows(gt(data.frame(a = c(1,2), b = NA_real_)), columns = vars(b),
                   fns = list(max = ~max(.)))
#> Error: Column `b` must be length 1 (the number of rows), not 0

Created on 2020-01-17 by the reprex package (v0.3.0)

[rich-iannone]

Thanks for reporting this and sorry for waiting until now to respond. I'm able to reproduce the error in my environment and I do believe this requires fixing.

[jkgrain]

I am running into the same issue. The following reprex might help trace the problem. I think it has to do with the mean of a vector that is all NA returning NaN even with na.rm=TRUE.

library(dplyr)
library(gt)

x <- c(1:6)
y <- c(6:2, NA)
z <- rep(NA, 6)

df <- tibble(x,y,z)

df <- mutate_if(df, is.logical, as.numeric)

df %>%
  gt() %>%
  grand_summary_rows(
    columns = c(1:3),
    fns = list(
      Avg = ~ mean(., na.rm = T)
    )
  ) 

mean(df$x, na.rm = TRUE)
mean(df$y, na.rm = TRUE)
mean(df$z, na.rm = TRUE)

[jkgrain]

I was able to trap for the problematic column with this variation on the code:

library(dplyr)
library(gt)

x <- c(1:6)
y <- c(6:2, NA)
z <- rep(NA, 6)

df <- tibble(x,y,z)

df %>%
  gt() %>%
  grand_summary_rows(
    columns = c(1:3),
    fns = list(
      Avg = ~ mean(., na.rm = T)
    )
  ) 

df %>%
  gt() %>%
  grand_summary_rows(
    columns = which(sapply(df, class) != "logical"),
    fns = list(
      Avg = ~ mean(., na.rm = T)
    )
  )