CODEWITHSUNDEEP
rstatalagequivalentlead
Jakob Jul Elben
Jakob Jul Elben

Reputation: 83

What's the R equivalent of Stata's _n function?

I am wondering whether there is a function in R like in Stata, where you are able to use the value of the observation n numbers before or after each observation. For instance if I need to multiply or divide with the observation before I would write something like this variable_x/variable_x[_n-1]

Upvotes: 3

Views: 8881

Answers (3)

Stefano Cacciatore
Stefano Cacciatore

Reputation: 45

The shortest way is:

bysort vect:gen n=_n 

vect=c(1,1,1,2,2,2,2,3,3,3,3,3,4)
n=ave(1:length(vect), vect, FUN = seq_along)

bysort vect:gen N=_N

ave(1:length(vect), vect, FUN = length)

Upvotes: 0

IRTFM
IRTFM

Reputation: 263451

I'm not exactly sure what the phrase "n numbers before" actually means. If it is an index then I may have interpreted it incorrectly. You can get the last value calculated with the semi-hidden .Last.value variable:

> for( i in 1:10) x <- x*.Last.value
> x
[1] 177147

> 3^11
[1] 177147

If you are using an index i to refer to an item, obj then obviously you could refer to obj[i-10]. There is an embed function that constructs a matrix of columns that are "shifted".

x <- 1:10
embed (x, 3)
     [,1] [,2] [,3]
[1,]    3    2    1
[2,]    4    3    2
[3,]    5    4    3
[4,]    6    5    4
[5,]    7    6    5
[6,]    8    7    6
[7,]    9    8    7
[8,]   10    9    8

Upvotes: 0

Frank
Frank

Reputation: 66819

In general, you can't get the exact same functionality. For example, in Stata, you might iterate with _n like...

clear
set obs 5
gen x = 1
replace x = x[ _n - 1 ]*1.1 if _n > 1
list 

     +--------+
     |      x |
     |--------|
  1. |      1 |
  2. |    1.1 |
  3. |   1.21 |
  4. |  1.331 |
  5. | 1.4641 |
     +--------+

In R, you can handle this case with the cumprod function. In other cases, you can use cumsum. And in others, lag (as mentioned by @Khashaa). These solutions cover most but not all cases.

If you really need to iterate using the prior row and can't use one of these shortcuts, you can still use a loop (with R syntax being similar to Stata's). If the loop is slow, you can also write it in C++ with the Rcpp package.

Upvotes: 1

Related Questions