how to delete dataframe's row with 3 of 5 equal element? - r

Question

I have a dataframe with 5 columns and many many rows, that have repetition of elements only for the first 3 columns (in short, it is a volume built by several volumes, and so there are same coordinates (x,y,z) with different labels, and I would like to eliminate the repeated coordinates).
How can I eliminate these with R commands?
Thanks
AV

Answer 1

You can use duplicated function, e.g. :
# create an example data.frame
Lab1<-letters[1:10]
Lab2<-LETTERS[1:10]
x <- c(3,4,3,3,4,2,4,3,9,0)
y <- c(3,4,3,5,4,2,1,5,7,2)
z <- c(8,7,8,8,4,3,1,8,6,3)
DF <- data.frame(Lab1,Lab2,x,y,z)
> DF
Lab1 Lab2 x y z
1 a A 3 3 8
2 b B 4 4 7
3 c C 3 3 8
4 d D 3 5 8
5 e E 4 4 4
6 f F 2 2 3
7 g G 4 1 1
8 h H 3 5 8
9 i I 9 7 6
10 j J 0 2 3
# remove rows having repeated x,y,z
DF2 <- DF[!duplicated(DF[,c('x','y','z')]),]
> DF2
Lab1 Lab2 x y z
1 a A 3 3 8
2 b B 4 4 7
4 d D 3 5 8
5 e E 4 4 4
6 f F 2 2 3
7 g G 4 1 1
9 i I 9 7 6
10 j J 0 2 3
EDIT :
To allow choosing amongst the rows having the same coordinates, you can use for example by function (even if is less efficient then previous approach) :
res <- by(DF,
INDICES=paste(DF$x,DF$y,DF$z,sep='|'),
FUN=function(equalRows){
# equalRows is a data.frame with the rows having the same x,y,z
# for exampel here we choose the first row ordering by Lab1 then Lab2
row <- equalRows[order(equalRows$Lab1,equalRows$Lab2),][1,]
return(row)
})
DF2 <- do.call(rbind.data.frame,res)
> DF2
Lab1 Lab2 x y z
0|2|3 j J 0 2 3
2|2|3 f F 2 2 3
3|3|8 a A 3 3 8
3|5|8 d D 3 5 8
4|1|1 g G 4 1 1
4|4|4 e E 4 4 4
4|4|7 b B 4 4 7
9|7|6 i I 9 7 6