Skip to content

filter data set unique, duplicate..

3 messages · Anders Bjørgesæter, Dimitris Rizopoulos, Sundar Dorai-Raj

Original
Anders Bjørgesæter
# 
Hello

First, thanks for the help for an earlier question about error handling!

I have problem filtering a dataset.
I'm trying to filter the data in the y columns based on the values in the x 
column, e.g.:

x          y1        y2                    yn
1.0       1          NA                  3
2.0       1          NA                  11
2.0       2          NA                  NA
3.0       1          5                      16
3.0       7          5                      2
4.0       8          4                      1

and want to keep the highest y if x is identical, like this:

x          y1        y2                    yn
1.0       1          NA                  3
2.0       2          NA                  11
3.0       7          5                      16
4.0       8          4                      1

or just as good:

x          y1        y2                    yn
1.0	   1          NA                  3
2.0       NA*    NA                  NA
2.0       2          NA                  11
3.0       NA*    5                      16
3.0       7          NA*                NA*
4.0       8          4                      1

If any has any suggestions or pointers how to do this I would really 
appreciate it.

/Anders
Dimitris Rizopoulos
# 
maybe you could consider something like this:

dat <- data.frame(x = c(1, 2, 2, 3, 3, 4),
                  y1 = c(1, 1, 2, 1, 7, 8),
                  y2 = c(NA, NA, NA, 5, 5, 4),
                  y3 = c(3, 11, NA, 16, 2, 1))
#############
out <- as.data.frame(lapply(dat[-1], function(y, x) tapply(y, x, max, 
na.rm = TRUE), x = dat["x"]))
out[out == -Inf] <- NA
out$x <- unique(dat["x"])
out


I hope it helps.

Best,
Dimitris

----
Dimitris Rizopoulos
Ph.D. Student
Biostatistical Centre
School of Public Health
Catholic University of Leuven

Address: Kapucijnenvoer 35, Leuven, Belgium
Tel: +32/16/336899
Fax: +32/16/337015
Web: http://www.med.kuleuven.be/biostat/
     http://www.student.kuleuven.be/~m0390867/dimitris.htm


----- Original Message ----- 
From: "Anders Bj??rges??ter" <anders.bjorgesater at bio.uio.no>
To: <r-help at stat.math.ethz.ch>
Sent: Wednesday, August 03, 2005 10:40 AM
Subject: [R] filter data set unique, duplicate..
Sundar Dorai-Raj
# 
Hi, Anders/Dimitris,
Dimitris Rizopoulos wrote:
Beware this line. If "x" is not sorted as it is in "dat" then your rows 
will be misaligned.

Here's another solution using "by" though it's no more efficient than 
what Dimitris has given.

out <- by(dat[-1], dat[1], function(y) {
   max.na <- function(x)
     if(all(is.na(x))) NA else max(x, na.rm = TRUE)
   apply(y, 2, max.na)
})
out <- as.data.frame(do.call("rbind", out))
out <- cbind(x = as.numeric(row.names(out)), out)
out

HTH,

--sundar