Remove all duplicates except last instance

Wouldn't this just be the standard case for using the 'fromLast' parameter to duplicated?

 dat[ !duplicated(dat[, c("ID", "Date")], fromLast=T),]
#---------
  ID     Date Tally
6  1 2/1/2011     6
7  3 2/1/2011     7
8  4 2/1/2011     8
9  2 2/1/2011     9

Your example was not rich enough to tell whether you needed the "Date" column in the test fro duplication, so perhaps you could simplify. I'm leaving it in to illustrate that duplicated has a data.frame method. I prefer !duplicated to unique because it allows easy access to the set complement if you are comparing groups.

Use dplyr:

data <- data %>%
  group_by(ID) %>%
  slice(which.max(Tally))

Use !rev(duplicated(rev(ID))) to filter out all but the last unique occurrences.

To get the dataset filtered, use dataset[!rev(duplicated(rev(dataset$ID))),]

Using a data.table join, you can set mult = 'last'

For example

library(data.table)
DT <- data.table(DF, key = 'id')

# join with the unique ID values
DT[unique(DT[,list(ID)]), mult= 'last']

   ID     Date Tally
1:  1 2/1/2011     6
2:  2 2/1/2011     9
3:  3 2/1/2011     7
4:  4 2/1/2011     8

If you knew the unique IDs you could also any of the following

DT[.(1:4), mult='last']
DT[list(1:4), mult = 'last']