如何将列中的特定值（如7）更改为R中的NA？

Question

我正在做一个项目，在我的专栏中，那些不知道的记录为7，而拒绝回答的记录为9。我试图找到一种简便的方法将所有这些值转换为NA。

Answer 1

You can use simple logic to assign NA, as in the following very simple example.

column <- c(1,2,1,3,7,7,1,2,9)
column[column %in% c(7,9)] <- NA

> column
[1]  1  2  1  3 NA NA  1  2 NA

Basically, you can operate on any column of a data frame using the $ operator. This treats the column as a vector. Using a logical operator on the vector returns a vector of TRUE or FALSE back, which can then be used to select the elements of the vector to change to NA.

I would caution that you might not always want to do that. Using NA instead of values in R can have annoying side effects because any operation against NA generally returns NA.

EDITED TO ADD: Per Gregor , I should mention that a data frame allows you to select single columns out of it using the $ operator. For instance, in data frame df.example , if the columns are A, B, and C, then df.example$A will extract column A as a vector. In contrast, the [ operator does not create vector subsets, and is used to select multiple columns of a data frame as a smaller data frame. For instance, given our example data frame, you could select columns A and B as a different data frame using df.example[c("A","B")] . For more guidance, trying running help('[') in R.

Answer 2

You can also use the replace() function to do this:

df <- data.frame(value = c(1, 3, 7, 3, 9, 9, 7, 2))
df$value <- replace(df$value, which(df$value==7 | df$value==9), NA)

Result:

> df
  value
1     1
2     3
3    NA
4     3
5    NA
6    NA
7    NA
8     2

This looks cumbersome, but I've found it convenient in situations where the which piece gets messier. Then you do that part on another line and just point this one at the object you created there.