set_col_as_date {dataPreparation}

Set columns as POSIXct

Description

Set as POSIXct a character column (or a list of columns) from a data.table.

Usage

set_col_as_date(data_set, cols = NULL, format = NULL, verbose = TRUE)

Arguments

data_set

Matrix, data.frame or data.table

cols

List of column(s) name(s) of data_set to transform into dates

format

Date's format (function will be faster if the format is provided) (character or list of character, default to NULL).
For timestamps, format need to be provided ("s" or "ms" or second or millisecond timestamps)

verbose

Should the function log (logical, default to TRUE)

Details

set_col_as_date is way faster when format is provided. If you want to identify dates and format automatically, have a look to identify_dates.
If input column is a factor, it will be returned as a POSIXct column.
If cols is kept to default (NULL) set_col_as_date won't do anything.

Value

data_set (as a data.table), with specified columns set as Date. If the transformation generated only NA, the column is set back to its original value.

Examples

# Lets build a data_set set
data_set <- data.frame(ID = seq_len(5),
                  date1 = c("2015-01-01", "2016-01-01", "2015-09-01", "2015-03-01", "2015-01-31"),
                  date2 = c("2015_01_01", "2016_01_01", "2015_09_01", "2015_03_01", "2015_01_31")
                  )

# Using set_col_as_date for date2
data_transformed <- set_col_as_date(data_set, cols = "date2", format = "%Y_%m_%d")

# Control the results
lapply(data_transformed, class)

# With multiple formats:
data_transformed <- set_col_as_date(data_set, format = list(date1 = "%Y-%m-%d", date2 = "%Y_%m_%d"))
lapply(data_transformed, class)

# It also works with timestamps
data_set <- data.frame(time_stamp = c(1483225200, 1485990000, 1488495600))
set_col_as_date(data_set, cols = "time_stamp", format = "s")