data.table通过另一个条件中的一个条件选择行

Question

I have a given datatable

A  B
1  2
3  4
3  2
2  1

I would like to select the row with the maximum A. If there is more than one row with the maximum A, I will select the row with the maximum B from the selected rows. Here should get

A  B
3  4

How should I achieve it using data.table ?

Answer 1

Data:

x <- read.table(header = TRUE, stringsAsFactors = FALSE, text = "
A  B
1  2
3  4
3  2
2  1")

Working code:

library(data.table)
x[ head(order(A, B, decreasing = TRUE), n = 1), ]
#    A B
# 1: 3 4

An alternative:

x[ order(A, B, decreasing = TRUE)[1], ]

is valid, but produces a row of NA when nrow(x)==0 , where I would prefer/expect 0 rows. For example:

x[0,][ order(A, B, decreasing = TRUE)[1], ]
#     A  B
# 1: NA NA
x[0,][ head(order(A, B, decreasing = TRUE), n = 1), ]
# Empty data.table (0 rows) of 2 cols: A,B

Another alternative, my first suggestion:

x[ order(A, B, decreasing = TRUE), ][1,]

It is perfectly legitimate but as @thelatemail suggested (thanks!), it is far less efficient in that it produces the whole (restructured) frame before giving you the first row.

Yet another alternative (from the crowd, thanks again thelatemail):

setorder(x, -A, -B)[1]