R data.table当前和所有先前行的最大值

Question

I have a data.table named "sampleDT" with column c1 and c2

require(data.table)
sampleDT <- data.table(c1 = c(1,2,3), c2 = c(7,5,6), result = c(NA, 7, 7))
print(sampleDT)
> print(sampleDT)enter code here
   c1 c2 result
1:  1  7     NA
2:  2  5      7
3:  3  6      7

I would like to create a new column 'result'; and the value is to take the maximum from previous row:

• 1st row of column 'result' would inevitably be NA;
• 2nd row would give 7;
• 3rd row would also give 7, because sampleDT[2,result] == 7 .

We can obviously use for-loop to achieve it; I'm wondering if there's any faster solution? (Note: not sure if shift would work, because result is dependent on the previous row.

Answer 1

You can leverage data.table 's chaining:

# load package
require(data.table)

# create dummy data
sampleDT <- data.table(c1 = c(1,2,3), c2 = c(7,5,6))

# run code
sampleDT[, result := shift(apply(.SD, 1, max)), 
           .SDcols = 1:2][, result := shift(apply(.SD, 1, max, na.rm = TRUE)), 
                            .SDcols = 1:3][]

>
   c1 c2 result
1:  1  7     NA
2:  2  5      7
3:  3  6      7