R中重复行中的最大值

Question

I have a data set including several variables such as ID,Time,Age, v1,v 2, v3. I need to replace the duplicated rows (condition on ID and time); for the rows which have the same ID and time, get the max for each variable and replace it in the data set (I need to keep all the duplicated rows). Any advice would be appreciated.

Answer 1

Using dplyr

library(dplyr)
your_data %>%
  group_by(ID, time) %>%
  mutate_each(funs = funs(max)))

If you have NA values, try

your_data %>%
  group_by(ID, time) %>%
  mutate_each(funs = funs(max(., na.rm = T)))

Happy to test/demo on any data provided.

Answer 2

My example has only v1 and v2, but you get the idea...

> head(d)
  ID Time Age v1 v2
1  a    1  11 12 13
2  a    1  21 12 53
3  a    1  11 42 43
4  b    1   4  6  7
5  b    2   1  2  3

Here we aggregate to get the max

> agg = aggregate(
          list(Age=d$Age,v1=d$v1,v2=d$v2),
          by=list(ID=d$ID,Time=d$Time),
          FUN=max
        )

> head(agg)
  ID Time Age v1 v2
1  a    1  21 42 53
2  b    1   4  6  7
3  b    2   1  2  3

Now we merge that with the first two columns of our original data

> merge(d[,c(1,2)],agg,by=c("ID","Time"))
  ID Time Age v1 v2
1  a    1  21 42 53
2  a    1  21 42 53
3  a    1  21 42 53
4  b    1   4  6  7
5  b    2   1  2  3

Answer 3

Another alternative using ave , as applied to @Larsenal's example data:

idvars <- c("ID","Time")
numvars <- setdiff(names(dat), idvars)
dat[numvars] <- lapply(dat[numvars], function(x) ave(x, dat[idvars], FUN=max))

#  ID Time Age v1 v2
#1  a    1  21 42 53
#2  a    1  21 42 53
#3  a    1  21 42 53
#4  b    1   4  6  7
#5  b    2   1  2  3