如何删除 R 中具有分类值的多个列？

Question

I know how to drop columns by name, but I am not quite sure how I am going to drop the columns which has categorical values. It can be done manually looking at which columns has categorical values, but not intuitive for me using R code. How to detect columns which has categorical values? any way to make this happen?

minimal data

mydf=structure(list(taken_time = c(15L, 5L, 39L, -21L, 46L, 121L), 
    ap6xl = c(203.2893857, 4.858269406, 2, 14220, 218.2215352, 
    115.5227706), pct5 = c(732.074484, 25.67901235, 1.01, 120.0477168, 
    3621.328567, 79.30561111), crp4 = c(196115424.7, 1073624.455, 
    1.23, 1457496.474, 10343851.7, 81288042.73), age = c(52L, 
    74L, 52L, 67L, 82L, 67L), gender = structure(c(2L, 2L, 2L, 
    1L, 2L, 1L), .Label = c("F", "M"), class = "factor"), inpatient_readmission_time_rtd = c(79.78819444, 
    57.59068053, 57.59068053, 57.59068053, 57.59068053, 9.893055556
    ), infection_flag = c(0L, 0L, 1L, 1L, 0L, 1L), temperature_value = c(98.9, 
    98.9, 98, 101.3, 99.5, 98.1), heartrate_value = c(106, 61, 
    78, 91, 120, 68), pH_result_time_rta = c(11, 85.50402145, 
    85.50402145, 85.50402145, 85.50402145, 85.50402145), gcst_value = c(15, 
    15, 15, 14.63769293, 15, 14.63769293)), row.names = c(NA, 
6L), class = "data.frame")

instead of manually typing name of columns which has categorical values, is there any way we can detect categorical columns and drop it?

I am concerning the case such as dataframe might have more than 10 categorical columns, it is sort of pain, so I am curious if it is possible using R. any thought?

for example, I can do this for above dataframe by manually looking at which one are categorical columns:

mydf <- mydf[!names(mydf) %in% c("gender", "infection_flag")]

is there any way we can detect which ones is categorical columns and drop it for numerical calculation purpose? any idea?

Answer 1

You can use dplyr and select all the numerical columns:

library(dplyr)

mydf %>% select_if(is.numeric)

Answer 2

An option with base R

i1 <- sapply(mydf, is.numeric)
df[i1]