Lag a 'timeSeries' object
stats-lag.Rd
Compute a lagged version of a "timeSeries"
object.
Arguments
- x
an object of class
timeSeries
.- k
-
an integer number, the number of lags (in units of observations). By default 1. Can also be a vector, in which case the result is a multivariate
"timeSeries"
in which columni
contains the series lagged byk[i]
, see the examples. - trim
a logical value. By default
TRUE
, the first missing observation in the return series will be removed.- units
-
an optional character string, which allows to overwrite the current column names of a
"timeSeries"
object. By defaultNULL
which means that the column names are selected automatically. - ...
arguments passed to other methods.
Examples
## Load Micsrosoft Data Set
x <- MSFT[1:20, "Open"]
## Lag the 'timeSeries' Object
lag(x, k = -1:1)
#> GMT
#> Open[-1] Open[0] Open[1]
#> 2000-09-27 60.8125 63.4375 NA
#> 2000-09-28 61.0000 60.8125 63.4375
#> 2000-09-29 60.5000 61.0000 60.8125
#> 2000-10-02 59.5625 60.5000 61.0000
#> 2000-10-03 56.3750 59.5625 60.5000
#> 2000-10-04 55.5000 56.3750 59.5625
#> 2000-10-05 55.8125 55.5000 56.3750
#> 2000-10-06 55.6250 55.8125 55.5000
#> 2000-10-09 53.9375 55.6250 55.8125
#> 2000-10-10 54.0000 53.9375 55.6250
#> 2000-10-11 56.3125 54.0000 53.9375
#> 2000-10-12 53.8750 56.3125 54.0000
#> 2000-10-13 53.5000 53.8750 56.3125
#> 2000-10-16 51.8750 53.5000 53.8750
#> 2000-10-17 49.6250 51.8750 53.5000
#> 2000-10-18 58.4375 49.6250 51.8750
#> 2000-10-19 61.3125 58.4375 49.6250
#> 2000-10-20 64.6250 61.3125 58.4375
#> 2000-10-23 62.6250 64.6250 61.3125
#> 2000-10-24 NA 62.6250 64.6250