n is an integer. The sequence I would like is:
1:n, 1:(n-1), 1:(n-2), ... , 1:3, 1:2, 1
Editor note:
In R, 1:n-1 is different from 1:(n-1). Be careful.
As short as sequence(n:1).
sequence(4:1)
#[1] 1 2 3 4 1 2 3 1 2 1
The function does no mystery (see note at the bottom):
function (nvec)
unlist(lapply(nvec, seq_len))
So snoram's answer reinvents the wheel. But calling seq_len is faster than calling other options as it is a primitive function (a C function) with only one argument.
sequence2 <- function (nvec) unlist(lapply(nvec, seq.int, from = 1))
sequence3 <- function (nvec) unlist(lapply(nvec, function(x) 1:x))
sequence4 <- function (nvec) unlist(lapply(nvec, seq.default, from = 1))
sequence5 <- function (nvec) unlist(lapply(nvec, seq, from = 1))
library(microbenchmark)
microbenchmark(sequence(100:1), sequence2(100:1),
sequence3(100:1), sequence4(100:1), sequence5(100:1))
#Unit: microseconds
# expr min lq mean median uq max
# sequence(100:1) 93.292 160.9325 205.5617 173.1995 200.0005 1157.201
# sequence2(100:1) 117.625 226.2120 308.4929 248.1055 280.8625 5477.710
# sequence3(100:1) 126.289 233.7875 365.6455 268.0495 301.8860 8808.911
# sequence4(100:1) 606.301 1195.4795 1463.3400 1237.5580 1344.3145 9986.619
# sequence5(100:1) 944.099 1864.3920 2063.3712 1942.2240 2119.3930 8581.593
## speed comparison
seq < seq.default < function(x) 1:x < seq.int < seq_len
s3 generic normal light-weighted user primitive 1-argument primitive
A note on 2022-07-31
Function sequence is now more powerful and more efficient. It no longer has source code as I showed above. I don't know when it was upgraded, though. I can't find relevant description in R NEWS.