ts_lag: Compute a lagged version of timeseries
In chriszheng2016/zstmodelr: Stock Factor Models for China Stock Market Investment

Description Usage Arguments Details Value Methods (by class) See Also

Generic function to shift data of timeseries(except date) forward and backward according current timeline

ts_lag(ts_dataset, k = 1, trim = TRUE, ...)

## S3 method for class 'tbl_df'
ts_lag(
  ts_dataset,
  k = 1,
  trim = TRUE,
  ...,
  date_index_field = c("date"),
  key_fields = NULL,
  parallel = getOption("zstmodelr.common.parallel", TRUE)
)

## S3 method for class 'timeSeries'
ts_lag(ts_dataset, k = 1, trim = TRUE, ...)

`ts_dataset`	a timeseries of tibble/timeseries.
`k`	an integer value. The number of lags (in units of observations). By default 1. k > 0: shift forward, using earlier data as current data k = 0: don't shift, keeping original data k < 0: shift backward, using later data as current data see details for more info.
`trim`	A logic flag of whether to remove the first missing observation in the return series. Default TRUE,
`...`	Arguments passed to other methods.
`date_index_field`	Name of date index field of ts_df for resample, default 'date', Column must be date-like. Only be used for tibble dataset.
`key_fields`	A character vector of key fields, which identify unique observation in each date. Only be used for tibble dataset.
`parallel`	A logic to determine whether to use parallel processing. default TRUE means to use parallel processing.

There are two types of lag operation:

shift forward: mostly known as lag, i.e. move data to next k periods, which means we use earlier data as current data while keeping current timeline
shift backward: mostly known as head, i.e. move previous k periods, which means use later data as current data while keeping current timeline