> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dqlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Use Cases

> Real-world Comparison metric examples in Prizm — data migration validation, staging vs. production parity, and pipeline checkpoint reconciliation.

<script type="application/ld+json">
  {`{
            "@context": "https://schema.org",
            "@type": "TechArticle",
            "headline": "Comparison Metric Use Cases",
            "description": "Real-world Comparison metric examples in Prizm - data migration validation, staging vs. production parity, and pipeline checkpoint reconciliation.",
            "url": "https://docs.dqlabs.ai/architecture/metrics/comparison/usecases",
            "publisher": {
              "@type": "Organization",
              "name": "DQLabs Inc",
              "logo": "https://media.brand.dev/332adc35-5bc4-4d2b-bf78-256aa4a5e414.svg"
            }
            }`}
</script>

## Example use cases

**Data migration validation:** A warehouse migration moves data from an on-premise Oracle database to Snowflake. Before cutover, a Comparison metric validates that `COUNT(*)` and `SUM(revenue)` match between both systems within 0.01% tolerance — running on every incremental load during the migration window. Any drift above the tolerance pauses the cutover until the discrepancy is resolved.

**Staging vs. production parity check:** Before promoting a dbt model change from staging to production, a Comparison metric verifies that the output of the staging model matches the production model on key aggregates within expected tolerances — catching unintentional logic changes before they affect live dashboards.

**Pipeline checkpoint reconciliation:** A comparison between the raw ingestion layer and the curated mart validates that `COUNT(DISTINCT order_id)` is consistent at every pipeline stage (Source → Lake → Warehouse → Mart). A drop at any stage surfaces the transformation that dropped records.

<Note>
  Both datasets must be accessible to Prizm through configured source connections. For row/value-level diff mode, ensure the service account has `SELECT` access to both tables and that a common key column exists for joining.
</Note>

<Tip>
  Use Comparison metrics to validate data migrations continuously — not just at cutover. Configure them to run before, during, and after migration so Prizm tracks parity throughout the process and immediately alerts if it breaks.
</Tip>