To find the top-performing products in each category, we can use the Deduplicate step on our product sales data:

We configure the step to use Category as the deduplication key (to group products by category) and Sales Amount as the group sort key (to order products within each category from highest to lowest sales), then keep the first 3 rows from each category:

The output will look like this - notice that only the top 3 products from each category are kept:

Suppose you have a customer dataset with duplicate entries and you want to keep only one record per customer based on their email address:

Configuration: - Deduplication key columns: Email (A-Z) - Rows to keep: First rows - Number of rows to keep: 1

Output:

To keep the most recent record for each customer, use the Group sort key columns to sort by date:

Configuration: - Deduplication key columns: Email (A-Z) - Group sort key columns: Purchase Date (Newest to Oldest) - Rows to keep: First rows - Number of rows to keep: 1

Using the same input data as above, the output would be:

To keep only records that have no duplicates (truly unique rows), use the Unique rows option:

Configuration: - Deduplication key columns: Product ID (Smallest to Largest) - Rows to keep: Unique rows

Output:

Only products with IDs 101 and 103 are kept because product 102 appears twice.

If you leave the Deduplication key columns empty, the entire dataset is treated as one group:

Configuration: - Deduplication key columns: (empty) - Group sort key columns: Amount (Largest to Smallest) - Rows to keep: First rows - Number of rows to keep: 5

Output:

This keeps the 5 rows with the highest amounts from the entire dataset.