根据行的值删除列

Question

Given two data frames:

 C1<-c(3,4,4,4,5)
 C2<-c(3,7,3,4,5)
 C3<-c(5,6,3,7,4)
 DF<-data.frame(C1=C1,C2=C2,C3=C3)
 DF
   C1 C2 C3
1  3  3  5
2  4  7  6
3  4  3  3
4  4  4  7
5  5  5  4

and

V1<-c(3,2,2,4,5)
V2<-c(3,7,3,5,2)
V3<-c(5,2,5,7,5)
V4<-c(1,1,2,3,4)
V5<-c(1,2,6,7,5)
DF2<-data.frame(V1=V1,V2=V2,V3=V3,V4=V4,V5=V5)
DF2
  V1 V2 V3 V4 V5
1  3  3  5  1  1
2  2  7  2  1  2
3  2  3  5  2  6
4  4  5  7  3  7
5  5  2  5  4  5

Looking at each equivalent row in both data frames, there is a relationship between the value in C3 and the number of columns I want to drop in that same row in DF2.

The relationship between the value in C3 and the # of columns in DF2 to drop looks like this

If C3≥7 drop V5
If C3=6.0:6.9 drop V4 and up (so basically V5,V4)
If C3=5.0:5.9 drop V3 and up (so basically V5,V4,V3)
If C3=4.0:4.9 drop V2 and up (so basically V5,V4,V3,V2)
If C3≤3.9 drop entire row

For this example, based on the values of C3, I would want DF2 to look like this

  V1 V2 V3 V4 V5
1  3  3
2  2  7  2

4  4  5  7  3 
5  5

I've tried write a simple script to do this (I'm pretty new so I like to keep things simple so I can see what's going on) but I'm throwing errors left and right so I'd appreciate some advice on how to proceed

Answer 1

Perhaps the easiest way is like:

DF3 <- DF2
for (i in seq_len(nrow(DF3))) {
  DF3[i, seq_len(ncol(DF3)) >= DF[i, ]$C3 - 2] <- NA
}
DF3

then,

> DF3
  V1 V2 V3 V4 V5
1  3  3 NA NA NA
2  2  7  2 NA NA
3 NA NA NA NA NA
4  4  5  7  3 NA
5  5 NA NA NA NA

Answer 2

I like Koshke's answer, but if your rules for setting to NA don't have a nice mathematical property to them or you need to define your rules arbitrarily, this approach should give you that flexibility. First, define a function that returns the columns to drop based on your rules:

f <- function(x) {
  if(x >= 7){
    out <- 5
  }else if(x >= 6.0){
           out <- 4:5
  } else if( x >= 5.0){
    out <- 3:5
  } else if (x >= 4.0){
    out <- 2:5
  } else {
    out <- 1:5
  }
  return(out)
}

Next, create a list for the column indices to drop:

z <- lapply(DF$C3, f)

Finally, loop through each row setting the corresponding columns to NA:

for(j in seq(length(z))){
  DF2[j, z[[j]]] <- NA
}

#-----
  V1 V2 V3 V4 V5
1  3  3 NA NA NA
2  2  7  2 NA NA
3 NA NA NA NA NA
4  4  5  7  3 NA
5  5 NA NA NA NA

Answer 3

A slight variation on kohske's answer using defined cut points:

breaksx <- cut(DF$C3,c(0,3,4,5,6,7,Inf),labels=FALSE)
for (i in seq(nrow(DF2))) {
        DF2[i,breaksx[i]:ncol(DF2)] <- NA
}

Result:

> DF2
  V1 V2 V3 V4 V5
1  3  3 NA NA NA
2  2  7  2 NA NA
3 NA NA NA NA NA
4  4  5  7  3 NA
5  5 NA NA NA NA

To remove the rows which are all NAs

DF2[apply(DF2,1,function(x) !all(is.na(x))),]

Result:

  V1 V2 V3 V4 V5
1  3  3 NA NA NA
2  2  7  2 NA NA
4  4  5  7  3 NA
5  5 NA NA NA NA