DATEFORMAT Function

• If the source Datetime value does not include a valid input for this function, a missing value is returned.
• Designer Cloud Enterprise Edition supports a wide variety of formats for Datetime fields. For more information on supported date formats, see Supported Data Types.
• You can explore the available Datetime formats through the Transformer page. From a column's type drop-down, select Date/Time. Then, select the formatting category. From the displayed drop-down, you can select a specific format. When this transform step is added to your recipe, you can edit it to see how the format is specified in Wrangle.

For more information on formatting numeric types, see NUMFORMAT Function.

Basic Usage

derive type:single value: DATEFORMAT(MyDate, 'yyyy-MM-dd') as: 'dateformat_MyDate'

Output: Generates a column of the valid date values in the MyDate column converted to year-month-day format.

Syntax and Arguments

derive type:single value:DATEFORMAT(Datetime_col, date_format_string)

For more information on syntax standards, see Language Documentation Syntax Notes.

datetime_col

Name of the column whose date data is to be formatted.

• Missing values for this function in the source data result in missing values in the output.
• Multiple columns and wildcards are not supported.

Usage Notes:

date_format_string

String value indicating the date format to apply to the input values. 

Designer Cloud Enterprise Edition supports Java formatting strings, with some exceptions. 

For more information on supported date formats, see Supported Data Types.

• Missing values for this function in the source data result in missing values in the output.
• Multiple columns and wildcards are not supported.

Usage Notes:

Examples

Example - formatting date values

This example illustrates several ways of wrangling heterogeneous date values, including the use of the DATEFORMAT function.

Source:

Your dataset includes the following messy date values:

Transform:

When this data is loaded into the application, it is not immediately recognized as a Datetime column, as the variation among the data complicates deciding on the proper date format. The first three rows look to be in a consistent format, but the other two are problematic.

You can try to change the column to a Datetime type with a format that matches the first three rows. You can select the appropriate format through the type drop-down. When previewed, the transform looks like the following:

settype col: MyDate type: 'Datetime','mm-dd-yy hh:mm:ss','mm*dd*yy*HH:MM'

When the column is reformatted, you should notice that the last two values in the column are mismatched. In the column histogram, you can see that date ranges include the 1999 date in the third row, so the final row should work if it was a valid date. 

The 1894 value looks like an outlier value and could be removed:

delete row: MATCHES([MyDate], `12/20/1894`)

For the remaining 1999 row, you can delete it or use the following transforms to conform it to the other rows. Use the following transform to change the 13 month value to a 12:

replace col: MyDate on: `13/` with: '12\/' global: true

The following two transforms complete the cleanup steps:

replace col: MyDate on: `/1999` with: '\/99' global: true

replace col: MyDate on: `:#+:00` with: ':00' global: true

If you apply the original formatting step, all dates are valid:

settype col: MyDate type: 'Datetime','mm-dd-yy hh:mm:ss','mm*dd*yy*HH:MM'

Now, your Datetime column can be formatted as needed using the dateformat function. The following step generates a new column that contains year, month, and day information as a single numeric value:

derive type:single value:DATEFORMAT(MyDate, 'yyyyMMdd')

Results:

The final dataset should look like the following:

Example - Other date formatting variations

For more information on supported date formats, see Supported Data Types.