how do i define a variable for subsetting multiple columns of a data.table in i?

Question

A common task of mine is filtering (subsetting) datasets in the data.tables format. I want to subset rows in i in a complex sort of way with multiple column-specific boolean conditions. When I get a new dataset, it will have the same type of columns and I will want to filter them in the same way for all datasets.

To illustrate my task, let me first create an example data.table.

library(data.table)

dt <- data.table(a = seq(1,6), b = letters[seq(1,6)], c = rep(c(4,3,2)))

This yields

   a b c
1: 1 a 4
2: 2 b 3
3: 3 c 2
4: 4 d 4
5: 5 e 3
6: 6 f 2

. Suppose I want to apply the following filtering criteria to the columns:

 dt[b != 'd'][c < 4][a < 6]

yielding

 a b c
1: 2 b 3
2: 3 c 2
3: 5 e 3

. Is there a way to convert that filtering criteria into a variable so that I can just tag it onto the end of the data.table?

I tried

x <- [b != 'd'][c < 4][a < 6]
dt[x]

but this throws the error

Error: unexpected '[' in "x <- ["

. This would be great because I could update the filtering strategy by changing just the variable x and have this filter then apply to all data.tables.

You need i1 <- dt[, b != 'd' & c < 4 & a < 6]; dt[i1]

akrun
– akrun

2019-04-09 16:25:49 +00:00
Commented Apr 9, 2019 at 16:25 — akrun
– akrun, Commented Apr 9, 2019 at 16:25

akrun · Accepted Answer · 2019-04-09 16:27:49Z

2

If it is to applied on different dataset, quote the expression and evaluate it on each dataset

i1 <- quote(b != 'd' & c < 4 & a < 6)
dt[dt[, eval(i1)]]
#   a b c
#1: 2 b 3
#2: 3 c 2
#3: 5 e 3

answered Apr 9, 2019 at 16:27

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

chinsoon12 Over a year ago

and also: dt[eval(i1)]

Collectives™ on Stack Overflow

how do i define a variable for subsetting multiple columns of a data.table in i?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related