On the Nature of Null

A small demonstration on handling missing data in a stream of sensor readings. The choice of strategy for dealing with `null` can significantly alter—or entirely invalidate—the outcome.

1. Define Handling Strategy

Ignore (Filter)

Exclude any record with a null value from the calculation. Simple, but reduces dataset size.

Coalesce (Impute)

Replace null values with a specified default. Preserves dataset size but introduces artificial data.

Propagate (Error)

Treat any null as a critical error, halting the entire calculation. The safest but most restrictive approach.

Human-in-the-Loop Validation

Impute `null` with:

Source Data Stream: `core_temp_c`

Timestamp	Device ID	Value (°C)

2. Aggregated Result

Awaiting calculation...

Processing Log

Ready to process data.