如何从 select 列组中删除 na？

Question

Very novice R user here. I have a data set and want avoid reducing my data set by a signficant amount (if I use na.omit or complex.cases it deletes ALL of the rows that contain na's, which massively shrinks my data set). I only want to remove the na's in the columns that are directly relevant to my project. Lets say column 1 and column 2 are relevant. I've tried to use foo2 <- na.omit(foo1[-3:-4]) but I'm met with an error "Warning, in 3:4, numeric expression has 2 elements, only first will be used".

I'd like to go from this

       column 1    column 2   column 3   column 4

 1         NA         4           3         9
 2         5          NA          NA        10
 3         8          10          NA        4
 4         11         6           2         NA

To this

        column 1    column 2   column 3   column 4

    3      8          10          NA        4
    4      11         6           2         NA

So instead of removing every single row, it only removed row 1 and 2.

Thank you in advance.

Answer 1

We can use complete.cases to return a logical vector (TRUE/FALSE - corresponds to no NA in a row/any NA) on the subset of columns and use that as row index to subset the full dataset

df1[complete.cases(df1[c('column1', 'column2')]),]
#    column1 column2 column3 column4
#3       8      10      NA       4
#4      11       6       2      NA

---

na.omit applied on the subset of columns only return the data without any NA for that subset data and not on the full dataset.

data

df1 <- data.frame(column1=c(NA,5,8,11),column2=c(4,NA,10,6),
         column3=c(3,NA,NA,2),column4=c(9,10,4,NA))

Answer 2

With df defined as your dataframe, you could determine the rows where neither column1 nor column2 is NA, and then make your selection:

df <- data.frame(column1=c(NA,5,8,11),column2=c(4,NA,10,6),column3=c(3,NA,NA,2),column4=c(9,10,4,NA))

df[with(df, !is.na(column1) & !is.na(column2)),]
  column1 column2 column3 column4
3       8      10      NA       4
4      11       6       2      NA

Answer 3

You can use drop_na from the tidyr package:

library(tidyr) # alternatively, you can load it from the tidyverse package

df <- tibble(
  col1 = c(NA_real_, 5, 8, 11),
  col2 = c(4, NA_real_, 10, 6),
  col3 = c(3, NA_real_, NA_real_, 2),
  col4 = c(9, NA_real_, 4, NA_real_)
)

drop_na(df, col1, col2)

# # A tibble: 2 x 4
#    col1  col2  col3  col4
#   <dbl> <dbl> <dbl> <dbl>
# 1     8    10    NA     4
# 2    11     6     2    NA