R 子集 dataframe - 最大值和 NA

Question

I have the following test - dataframe

df <- data.frame(V1 = c(1, 2, 3), V2 = c(0, 5, NA), V3=c(NA, 10, NA), V4=c(2, 2, NA))    

> df

   V1 V2 V3 V4
 1  1  0 NA  2
 2  2  5 10  2
 3  3 NA NA NA       

Now I want to subset this dataframe:

• keep rows where the max value of V2:V4 is smaller than 3

OR

• keep rows where all values of V2:V4 are NA

So the result should look like this:

df_new

   V1 V2 V3 V4
 1  1  0 NA  2
 3  3 NA NA NA   

Only the first & third row of the original dataframe are kept.

I could use the following command:

subset(DF, (is.na(V2) & is.na(V3) & is.na(V4)) | ((V2 < 3 | is.na(V2)) & (V3 < 3 | is.na(V3)) & (V4 < 3 | is.na(V4))))

to do this. But it's quite tedious & in my real-life data frame has > 30 columns to check, so there must be a better way of doing this.

Answer 1

You can do:

df[rowSums(df[, 2:4] >= 3, na.rm = TRUE) == 0, ]

  V1 V2 V3 V4
1  1  0 NA  2
3  3 NA NA NA

Answer 2

In dplyr , we can use filter_at to select specific columns to check, replace NA values to 0 and select rows where all values are less than 3.

library(dplyr)
df %>% filter_at(vars(V2:V4), all_vars(replace(., is.na(.), 0) < 3))

#  V1 V2 V3 V4
#1  1  0 NA  2
#2  3 NA NA NA