CaseWhenMixed / Manipulation Layer

Create a new column using flexible conditional logic that supports both constant values and column references. Similar to a hybrid of SQL's CASE WHEN and pandas' np.select() with mixed literal and column values.

Common applications:

• Data standardization (map values against reference tables)
• Threshold-based classification (fixed bounds with dynamic values)
• Quality scoring (compare against standards and peer values)
• Alert generation (mix of absolute and relative triggers)
• Business rules implementation (fixed and dynamic criteria)

Example: Product Pricing Strategy

Implementation logic:

CASE 
  WHEN stock_level < 20 THEN base_price * 1.10             -- Constant multiplier
  WHEN competitor_price < base_price THEN competitor_price -- Column value
  WHEN stock_level > 100 THEN 'Standard'                   -- Constant text
  ELSE base_price                                          -- Column fallback
END

Key features:

• Mix constant values and column references in conditions
• Compare columns against fixed thresholds
• Use both static and dynamic replacement values
• Flexible type handling through AsDatatype
• Configurable null value handling

Note: When mixing types, ensure compatibility and use AsDatatype appropriately for the desired output format.