Skip to contents

Coerce time series objects and tibbles with date/date-time columns to xts.

Usage

tk_xts(data, select = NULL, date_var = NULL, silent = FALSE, ...)

tk_xts_(data, select = NULL, date_var = NULL, silent = FALSE, ...)

Arguments

data

A time-based tibble or time-series object.

select

Applicable to tibbles and data frames only. The column or set of columns to be coerced to ts class.

date_var

Applicable to tibbles and data frames only. Column name to be used to order.by. NULL by default. If NULL, function will find the date or date-time column.

silent

Used to toggle printing of messages and warnings.

...

Additional parameters to be passed to xts::xts(). Refer to xts::xts().

Value

Returns a xts object.

Details

tk_xts is a wrapper for xts::xts() that is designed to coerce tibble objects that have a "time-base" (meaning the values vary with time) to xts class objects. There are three main advantages:

  1. Non-numeric columns that are not removed via select are dropped and the user is warned. This prevents an error or coercion issue from occurring.

  2. The date column is auto-detected if not specified by date_var. This takes the effort off the user to assign a date vector during coercion.

  3. ts objects are automatically coerced if a "timetk index" is present. Refer to tk_ts().

The select argument can be used to select subsets of columns from the incoming data.frame. Only columns containing numeric data are coerced. The date_var can be used to specify the column with the date index. If date_var = NULL, the date / date-time column is interpreted. Optionally, the order.by argument from the underlying xts::xts() function can be used. The user must pass a vector of dates or date-times if order.by is used.

For non-data.frame object classes (e.g. xts, zoo, timeSeries, etc) the objects are coerced using xts::xts().

tk_xts_ is a nonstandard evaluation method.

Examples

library(dplyr)

### tibble to xts: Comparison between tk_xts() and xts::xts()
data_tbl <- tibble::tibble(
    date = seq.Date(as.Date("2016-01-01"), by = 1, length.out = 5),
    x    = rep("chr values", 5),
    y    = cumsum(1:5),
    z    = cumsum(11:15) * rnorm(1))

# xts: Character columns cause coercion issues; order.by must be passed a vector of dates
xts::xts(data_tbl[,-1], order.by = data_tbl$date)
#>                     x  y         z
#> 2016-01-01 chr values  1  2.060373
#> 2016-01-02 chr values  3  4.308053
#> 2016-01-03 chr values  6  6.743040
#> 2016-01-04 chr values 10  9.365333
#> 2016-01-05 chr values 15 12.174933

# tk_xts: Non-numeric columns automatically dropped; No need to specify date column
tk_xts(data_tbl)
#> Warning: Non-numeric columns being dropped: date, x
#> Using column `date` for date_var.
#>             y         z
#> 2016-01-01  1  2.060373
#> 2016-01-02  3  4.308053
#> 2016-01-03  6  6.743040
#> 2016-01-04 10  9.365333
#> 2016-01-05 15 12.174933

# ts can be coerced back to xts
data_tbl %>%
    tk_ts(start = 2016, freq = 365) %>%
    tk_xts()
#> Warning: Non-numeric columns being dropped: date, x
#> Warning: 'tzone' attributes are inconsistent
#>             y         z
#> 2016-01-01  1  2.060373
#> 2016-01-02  3  4.308053
#> 2016-01-03  6  6.743040
#> 2016-01-04 10  9.365333
#> 2016-01-05 15 12.174933

### Using select and date_var
tk_xts(data_tbl, select = y, date_var = date)
#>             y
#> 2016-01-01  1
#> 2016-01-02  3
#> 2016-01-03  6
#> 2016-01-04 10
#> 2016-01-05 15


### NSE: Enables programming
date_var <- "date"
select   <- "y"
tk_xts_(data_tbl, select = select, date_var = date_var)
#>             y
#> 2016-01-01  1
#> 2016-01-02  3
#> 2016-01-03  6
#> 2016-01-04 10
#> 2016-01-05 15