CaseWhen / Manipulation Layer

Create a new column by applying multiple conditional statements across one or more columns. Similar to SQL's CASE WHEN or pandas' np.select(). Evaluates conditions in sequence and assigns values based on the first matching condition.

Common applications:

• Order processing status
• Data quality validation
• Customer categorization
• Product classification
• Transaction flagging

Example: Order Processing Priority Assignment

Implementation logic:

CASE 
  WHEN shipping = 'Next Day' THEN 'Urgent'
  WHEN value > 300 THEN 'High'
  WHEN order_type = 'Express' THEN 'Medium'
  WHEN status = 'Complete' THEN 'Low'
  ELSE 'Normal'
END

Key points:

• Each condition evaluates a single column
• Conditions are checked in sequence
• First matching condition determines result
• Non-matching rows use the Otherwise value
• More specific conditions should be listed first

Note: The order of conditions is crucial as each row's value comes from the first matching condition. Consider business logic priority when ordering conditions.