R清洁数据框的聪明方法

Question

I have a data frame with two columns, an index column, which indexes rows in a second data frame. These rows all contain a specific event. Which event this is, is coded in the second column, here named code_start_stop .

Example:

index <- c(769, 766, 810, 813, 830, 842, 842, 892, 907, 944)
code_start_stop <- c(2006, 2001, 2004, 1001, 1004, 2001, 1001, 1006, 2004, 1004)
replace_all <- data.frame(index, code_start_stop)

Now there are pairs of start/stop codes, ie 2001 and 1001, 2002 and 1002, etc. The aim is that, in case there are rows that are enclosed by a start marker (ie here 2006) and the respective next stop marker (here 1006) , these rows should be removed from the data frame. Note that there are always pairs of start & stop markers.

Any suggestions for a clever way how to do this are appreciated. Thanks!

Answer 1

Your question is a little confusing, please correct me if I got it wrong. The following should work:

startm <- 2006 #startmarker
endm   <- 1006 #endmarker

 #look for  row that contains markers
 index1 <- which(replace_all[,2]  == startm) 
 index2 <- which(replace_all[,2]  == endm)

 #subset accordingly
 replace_all <- replace_all[-(index1:index2),]

Note: This removes also the rows, containing the markers. If you want to only remove rows between the markes add a +1/-1 at the subsetting step.

Answer 2

The solution is now based on maRtin's suggestion and seems to work pretty well.

I do the following going through all pairs of start and end markers:

to_delete <- c()
## Care = 2001/1001
startm1 <- 2001
endm1 <- 1001
index1 <- which((replace_all[,2]  == startm1))
index2 <- which((replace_all[,2]  == endm1))
if(length(index1) !=0){
  for (i in 1:length(index1)){
    if (index2[i]-index1[i]>1){
      to_delete <- c(to_delete, (((index1[i])+1):((index2[i])-1)))
    }
  }
}

... go through all other pairs of start/stop markers and then remove to_delete

if (length(to_delete) != 0){
    replace_all <- replace_all[-to_delete,]
  }
    replace_all <- replace_all[,1]
  }