Construct a validated time series — morie

Wraps stats::ts() with explicit validation: the input must be a strictly ordered, regular numeric series with no implicit missing periods. This is the single pre-processing routine through which all subsequent functions receive their data, so time attributes are maintained uniformly.

Usage

morie_ts(
  x,
  frequency = 1,
  start = 1,
  calendar = NA_character_,
  na_action = c("error", "interpolate", "omit"),
  units = attr(x, "units")
)

Arguments

x: A numeric vector (or an object coercible to one, including a units-typed vector, which is coerced via as.numeric()), or an existing ts.
frequency: Number of observations per unit of time (e.g. 12 for monthly-in-years). Default 1.
start: Time of the first observation. Default 1.
calendar: Character tag documenting the calendar system assumed (e.g. "monthly"); recorded as an attribute, not interpreted.
na_action: How to handle missing values (regular series require an explicit decision): "error" (default) rejects implicit missing values; "interpolate" fills them by linear interpolation (imputation); "omit" drops them with a warning (the series is then treated as irregular).
units: Optional character unit label carried as an attribute and propagated to return values.

Value

A morie_ts object (a ts with validated ordering).

Examples

morie_ts(cumsum(rnorm(24)), frequency = 12, start = 2000)
#>              Jan         Feb         Mar         Apr         May         Jun
#> 2000 -0.19819542  0.69381297  0.66809790  0.02043745  0.66679686  0.23296412
#> 2001 -1.75296930 -1.51616563 -3.85888875 -2.89719212 -3.50161785 -4.25449513
#>              Jul         Aug         Sep         Oct         Nov         Dec
#> 2000  2.00557531  1.98731560  2.84013059  3.04529350  0.03724490 -1.32886704
#> 2001 -5.81010672 -7.26400046 -7.20766863 -6.69829922 -8.79618218 -9.80054416