查找唯一行时忽略数据帧中的NA

Question

I have a dataframe with 20 columns and about 200 rows, and I would like to find the unique rows. The problem is that nearly every row has a few NAs mixed in: this is really missing data and I would like the NAs to be treated like a "wildcard", not used to match other NAs.

The following two rows should be recognized as a match (ie non-unique)

T, S, NA, Z
NA, S, G, Z

I've tried the incomparables argument to the unique function, but it doesn't seem to be implemented. Thanks a lot.

Answer 1

Put this in a double for loop:

all(na.omit(x[1,] == x[2,]))

Replacing 1 and 2 with i and j to cycle through all comparisons you need to check.

Answer 2

You could try

val <-  apply(df, 1, function(x) {paste(na.omit(x), collapse='')})
df[!duplicated(val),]
#    V1 V2   V3 V4
#1    T  S <NA>  Z
#2 <NA>  S    G  Z
#3    S  G    Z  T

data

 df <- structure(list(V1 = c("T", NA, "S", "S", "S"), V2 = c("S", "S", 
 "G", NA, "G"), V3 = c(NA, "G", "Z", "Z", NA), V4 = c("Z", "Z", 
 "T", "G", "Z")), .Names = c("V1", "V2", "V3", "V4"), row.names = c(NA, 
 -5L), class = "data.frame")