The R Project SVN R

% File src/library/stats/man/ts.Rd
% Part of the R package, https://www.R-project.org
% Copyright 1995-2025 R Core Team
% Distributed under GPL 2 or later

\name{ts}
\title{Time-Series Objects}
\alias{ts}
\alias{as.ts}
\alias{as.ts.default}
\alias{is.ts}
\alias{Ops.ts}
\alias{cbind.ts}
\alias{is.mts}
\alias{[.ts}
\alias{t.ts}
\alias{head.ts}
\alias{tail.ts}
\description{
  The function \code{ts} is used to create time-series objects.

  \code{as.ts} and \code{is.ts} coerce an object to a time-series and
  test whether an object is a time series.
}
\usage{
ts(data = NA, start = 1, end = numeric(), frequency = 1,
   deltat = 1, ts.eps = getOption("ts.eps"),
   class = if(nseries > 1) c("mts", "ts", "matrix", "array") else "ts",
   names = )
as.ts(x, \dots)
is.ts(x)

is.mts(x)
}
\arguments{
  \item{data}{a vector or matrix of the observed time-series
    values. A data frame will be coerced to a numeric matrix via
    \code{data.matrix}.  (See also \sQuote{Details}.)}
  \item{start}{the time of the first observation.  Either a single
    number or a vector of two numbers (the second of which is an integer),
    which specify a natural time
    unit and a (1-based) number of samples into the time unit.  See
    the examples for the use of the second form.}
  \item{end}{the time of the last observation, specified in the same way
    as \code{start}.}
  \item{frequency}{the number of observations per unit of time.}
  \item{deltat}{the fraction of the sampling period between successive
    observations; e.g., 1/12 for monthly data.  Only one of
    \code{frequency} or \code{deltat} should be provided.}
  \item{ts.eps}{time series comparison tolerance.  Frequencies are
    considered equal if their absolute difference is less than
    \code{ts.eps}.  It is also used to check consistency of \code{end - start},
    \code{frequency}, and the length of the time-series.}
  \item{class}{class to be given to the result, or none if \code{NULL}
    or \code{"none"}.  The default is \code{"ts"} for a single series, or
    \code{c("mts", "ts", "matrix", "array")} for multiple series.}
  \item{names}{a character vector of names for the series in a multiple
    series: defaults to the colnames of \code{data}, or \code{"Series 1"},
    \code{"Series 2"}, \dots.}
  \item{x}{an arbitrary \R object.}
  \item{\dots}{arguments passed to methods (unused for the default method).}
}
\details{
  The function \code{ts} is used to create time-series objects.  These
  are vectors or matrices which inherit from class \code{"ts"} (and have
  additional attributes) which represent data sampled at equispaced
  points in time.  In the matrix case, each column of the matrix
  \code{data} is assumed to contain a single (univariate) time series.
  Time series must have at least one observation, and although they need
  not be numeric there is very limited support for non-numeric series.

  Class \code{"ts"} has a number of methods.  In particular arithmetic
  will attempt to align time axes, and subsetting to extract subsets of
  series can be used (e.g., \code{EuStockMarkets[, "DAX"]}).  However,
  subsetting the first (or only) dimension will return a matrix or
  vector, as will matrix subsetting.  \I{Subassignment} can be used to
  replace values but not to extend a series (see \code{\link{window}}).
  There is a method for \code{\link{t}} that transposes the series as a
  matrix (a one-column matrix if a vector) and hence returns a result
  that does not inherit from class \code{"ts"}.

  Argument \code{frequency} indicates the sampling frequency of the
  time series, with the default value \code{1} indicating one sample in
  each unit time interval.  For
  example, one could use a value of \code{7} for \code{frequency} when
  the data are sampled daily, and the natural time period is a week, or
  \code{12} when the data are sampled monthly and the natural time
  period is a year.  Values of \code{4} and \code{12} are assumed in
  (e.g.) \code{print} methods to imply a quarterly and monthly series
  respectively.  \code{frequency} need not be a whole
  number:  for example, \code{frequency = 0.2} would imply sampling
  once every five time units.

  \code{as.ts} is generic.  Its default method will use the
  \code{\link{tsp}} attribute of the object if it has one to set the
  start and end times and frequency.

  \code{is.ts()} tests if an object is a time series, i.e., inherits from
  \code{"ts"} and is of positive length.

  \code{is.mts(x)} tests if an object \code{x} is a multivariate time series,
  i.e., fulfills \code{is.ts(x)}, \code{is.matrix(x)} and inherits from
  class \code{"mts"}.
}
\references{
  \bibshow{R:Becker+Chambers+Wilks:1988}
}
\seealso{
  \code{\link{tsp}},
  \code{\link{frequency}},
  \code{\link{start}},
  \code{\link{end}},
  \code{\link{time}},
  \code{\link{window}};
  \code{\link{print.ts}}, the print method for time series objects;
  \code{\link{plot.ts}}, the plot method for time series objects.

  For other definitions of \sQuote{time series} (e.g.,
  time-ordered observations) see the CRAN task view at
  \url{https://CRAN.R-project.org/view=TimeSeries}.
}
\examples{
require(graphics)

ts(1:10, frequency = 4, start = c(1959, 2)) # 2nd Quarter of 1959
print( ts(1:10, frequency = 7, start = c(12, 2)), calendar = TRUE)
# print.ts(.)
## Using July 1954 as start date:
gnp <- ts(cumsum(1 + round(rnorm(100), 2)),
          start = c(1954, 7), frequency = 12)
plot(gnp) # using 'plot.ts' for time-series plot

\donttest{utils:: methods(class = "ts") # all functions with methods available for "ts"}

## "ts" methods for head() and tail()
utils:: head(lynx, 4)
utils:: tail(lynx, -7)
\dontshow{ stopifnot(is.ts(head(lynx, -6)),
                     is.ts(tail(lynx, 3)))
}

## Multivariate
z <- ts(matrix(rnorm(300), 100, 3), start = c(1961, 1), frequency = 12)
class(z)
is.mts(z)
head(z) # "ts" method ==> incl. times:
plot(z)
plot(z, plot.type = "single", lty = 1:3)

## A phase plot:
plot(nhtemp, lag(nhtemp, 1), cex = .8, col = "blue",
     main = "Lag plot of New Haven temperatures")
}
\keyword{ts}