Keep Transform

keep row:(customerStatus == 'active')

Output: For each row in the dataset, if the value of the customerStatus column is active, then the row is retained. Otherwise, the row is deleted from the dataset.

keep row:(expression)

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

Expression to identify the row or rows on which to perform the transform. Expression must evaluate to true or false.

Examples:

Score >= 50

LEN(LastName) > 8

ISMISSING([Title])

ISMISMATCHED(Score,['Integer'])

For the keep transform, if the expression for the row parameter evaluates to true for a row, it is kept in the dataset. Otherwise, it is removed.

Example:

keep row: (lastOrder >= 10000 && status == 'Active')

Output: Retains all rows in the dataset where the lastOrder value is greater than or equal to 10,000 and the customer status is Active.

Usage Notes:

This examples illustrates how you can keep and delete rows from your dataset.

Source:

Your dataset includes the following order information. You want to edit your dataset so that:

Transformation:

First, you remove the orders for old products. Since the set of products is relatively small, you can start first by adding the following:

Custom formula

Custom single

(ProdId == 'P100')

Delete matching rows

When this step is previewed, you should notice that the top row in the above table is highlighted for removal. Notice how the transformation relies on the ProdId value. If you look at the ProductName value, you might notice that there is a misspelling in one of the affected rows, so that column is not a good one for comparison purposes.

You can add the other product IDs to the transformation in the following expansion of the transformation, in which any row that has a matching ProdId value is removed:

Custom formula

Custom single

(ProdId == 'P100' || ProdId == 'P101' || ProdId == 'P102' || ProdId == 'P103')

Delete matching rows

When the above step is added to your recipe, you should see data that looks like the following:

Now, you can filter out of the dataset orders that are older than 90 days. First, add a column with today's date:

Single row formula

'2/25/16'

'today'

Keep the rows that are within 90 days of this date using the following:

Custom formula

Custom single

datedif(OrderDate,today,day) <= 90

Keep matching rows

Don't forget to delete the today column, which is no longer needed:

today

Delete selected columns

Results: