Multiple Lags with Dplyr

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Multiple Lags with Dplyr

Lorenzo Isella
Dear All,
I refer to the excellent post at

https://purrple.cat/blog/2018/03/02/multiple-lags-with-tidy-evaluation/

What I want to do is to create a function capable, à la dplyr, to
generate new columns which are a lagged version of existing columns in
a data frame.
For instance, you can do this manually as


d2 <- tibble(x1 =1:10, x2=10:19,  x3=50:59)


d3 <- d2%>%mutate(x1lag1=lag(x1, 1), x1lag2=lag(x1,2))


but this becomes quickly tedious when you need to take several lags of
different columns.
One solution in the link above is the following


lags <- function(var, n=10){
  var <- enquo(var)
 
  indices <- seq_len(n)
  map( indices, ~quo(lag(!!var, !!.x)) ) %>%
    set_names(sprintf("lag_%s_%02d", quo_text(var), indices))
 
}


d4 <- d2 %>%
  mutate( !!!lags(x1, 3), !!!lags(x2,3) )


does anybody know how this could be made more general? I mean that I
would like to take a fixed number of lags of a list of columns (x1 and
x2, for instance), just by passing the list of columns and without
repeating the commands for x1 and x2.
Any suggestion is appreciated.
Cheers

Lorenzo

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: Multiple Lags with Dplyr

Gabor Grothendieck
lag.zoo supports vector-based lags on zoo objects.
A few caveats:

- dplyr's lag clobbers the base R lag (which you need to
invoke lag's methods) so if you have dplyr loaded be sure
to refer to stats::lag.

- dplyr's lag works backwards relative to the standard set
in base R so dplyr::lag(x, 1) corresponds to stat::lag(x, -1) in
base R

- zoo follows base R's standard

- you can use as.data.frame or fortify.zoo to convert a zoo
object to a data frame if you need that.  The first one
drops the time index and the second one includes it.

  library(zoo)
  stats::lag(zoo(d2$x1), 0:-2)

giving this zoo object:

   lag0 lag-1 lag-2
1     1    NA    NA
2     2     1    NA
3     3     2     1
4     4     3     2
5     5     4     3
6     6     5     4
7     7     6     5
8     8     7     6
9     9     8     7
10   10     9     8


On Tue, Apr 23, 2019 at 9:10 AM Lorenzo Isella <[hidden email]> wrote:

>
> Dear All,
> I refer to the excellent post at
>
> https://purrple.cat/blog/2018/03/02/multiple-lags-with-tidy-evaluation/
>
> What I want to do is to create a function capable, ą la dplyr, to
> generate new columns which are a lagged version of existing columns in
> a data frame.
> For instance, you can do this manually as
>
>
> d2 <- tibble(x1 =1:10, x2=10:19,  x3=50:59)
>
>
> d3 <- d2%>%mutate(x1lag1=lag(x1, 1), x1lag2=lag(x1,2))
>
>
> but this becomes quickly tedious when you need to take several lags of
> different columns.
> One solution in the link above is the following
>
>
> lags <- function(var, n=10){
>   var <- enquo(var)
>
>   indices <- seq_len(n)
>   map( indices, ~quo(lag(!!var, !!.x)) ) %>%
>     set_names(sprintf("lag_%s_%02d", quo_text(var), indices))
>
> }
>
>
> d4 <- d2 %>%
>   mutate( !!!lags(x1, 3), !!!lags(x2,3) )
>
>
> does anybody know how this could be made more general? I mean that I
> would like to take a fixed number of lags of a list of columns (x1 and
> x2, for instance), just by passing the list of columns and without
> repeating the commands for x1 and x2.
> Any suggestion is appreciated.
> Cheers
>
> Lorenzo
>
> ______________________________________________
> [hidden email] mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.



--
Statistics & Software Consulting
GKX Group, GKX Associates Inc.
tel: 1-877-GKX-GROUP
email: ggrothendieck at gmail.com

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.