> ## 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.
# Setup
> Step-by-step guide to creating a Conditional metric in Prizm, covering rule types, select criteria, and threshold configuration.
## Creating a conditional metric
1. Navigate to **Metric** in the top navigation.
2. Click the **Add** button (⊕).
3. Enter a **Metric Name**.
4. Select **Subcategory: Conditional** from the dropdown.
5. Enter a **Purpose** description.
6. Click **Create**.
Click the metric name to open its detail page, and provide the following details
* **Asset** — the table the metric evaluates
* **Metadata tabs** — Description, Purpose, Rationale
* **Conditional Metric Configuration** — three tabs: Configuration, Total Records, Failed Rows
* **Threshold**
* **Semantic context** and **Owners**
**Asset:** Select the table or view this metric will evaluate. Use the **AI-assisted asset selection** button to let Prizm suggest an asset based on the metric name and purpose.
The **Downstream Impact** counter shows how many other assets depend on this asset — useful for understanding blast radius.
Fill in the metadata tabs as needed:
| Tab | Content |
| --------------- | --------------------------------------------------------------------------------------- |
| **Description** | What this metric measures. Prizm AI can suggest a description based on the metric name. |
| **Purpose** | Business justification — why this metric exists. |
| **Rationale** | The reasoning behind the specific rule chosen. |
The **Configuration** tab is the core of a conditional metric. It has three sub-sections.
#### Select criteria
Expand **Select criteria** to scope which rows are evaluated. By default this uses `SELECT *` (all rows). Choose between:
* **All columns** — evaluate all columns in the asset
* **Selected columns** — restrict evaluation to specific columns
Leave Select criteria empty to evaluate every row in the asset.
#### Condition rules
Define what "valid" means for each row. Use the logical operators **NOT / AND / OR** to combine multiple rules. Click **+ Add rule** to add a rule or **+ Add group** to create a nested rule group.
Select a **Rule type** from the toolbar, then click **+ Add rule**. Each rule type exposes different fields:
**Per-row condition on a single column.** Validates that each record's column value meets a specific condition — e.g., `QTY_SOLD Between 10 and 20`.
1. Click **Row check** in the rule type toolbar.
2. Click **+ Add rule**.
3. **Select column** — choose the column to evaluate.
4. **Select operator** — choose the comparison (e.g., `=`, `!=`, `Between`, `Is Null`, `Is Not Null`).
5. For range operators (e.g., `Between`), enter a **Start value** and **End value**.
6. For single-value operators, enter the comparison **Value**.
**Compares one column to another.** Enforces cross-column consistency — e.g., `end_date >= start_date`.
1. Click **Col vs col** in the rule type toolbar.
2. Click **+ Add rule**.
3. **Select column** — choose the first (left-hand) column.
4. **Select operator** — choose the comparison operator.
5. A second column picker appears — select the **comparison column**.
**Set-level assertion — evaluates across all rows.** Enforces dataset-wide constraints — e.g., no duplicate values in a column.
1. Click **Aggregate** in the rule type toolbar.
2. Click **+ Add rule**.
3. **Function** — select the aggregate function from the dropdown (e.g., `No duplicates on`).
4. **Select columns** — choose one or more columns the function applies to.
**Conditional logic — condition only fires when the guard clause is true.** Use when a rule should apply only to rows that satisfy a prior condition — e.g., if `order_type = 'express'` then `carrier_tracking_number` is not null.
1. Click **If–then** in the rule type toolbar.
2. Click **+ Add rule**.
3. In the **IF** section — define the guard clause:
* **Select column** — the guard column.
* **Select operator** — the guard condition.
4. In the **THEN** section — define the required outcome:
* **Select column** — the column that must satisfy the consequence.
* **Select operator** — the consequence condition.
**Build a WHERE predicate — rows must satisfy this condition.** Use for complex expressions that cannot be captured by the other rule types. The expression builder combines columns, functions, and operators into a custom SQL WHERE clause.
1. Click **Expression** in the rule type toolbar.
2. Click **+ Add rule**.
3. In the **COLUMNS** panel on the left, click a column name to insert it into the expression area.
4. Switch to the **FUNCTIONS** tab to insert SQL functions.
5. Use the operator buttons to build the full condition: `AND`, `OR`, `NOT`, `=`, `!=`, `<`, `>`, `<=`, `>=`, `BETWEEN`, `IN`, `IS NULL`, `IS NOT NULL`.
6. The completed expression becomes the WHERE predicate evaluated row-by-row. Rows satisfying it are counted as valid.
#### Rule options
| Option | Description |
| ----------------- | ----------------------------------------------------------------- |
| **Valid** | Count only rows that satisfy the rule as valid records |
| **Scoring** | Include this rule in the asset quality score |
| **Weightage%** | Relative weight of this metric in the overall score (default 100) |
| **Pass Criteria** | Score percentage above which the rule passes (default `>= 100`) |
| **Export** | Include rule results in data exports |
Click **Validate** to test the rule against current data and preview Valid / Invalid counts. Click **View prepared query** to inspect the underlying SQL Prizm generates.
The **Total Records** and **Failed Rows** tabs let you override how Prizm counts the denominator and numerator of the score — useful when the default row counts don't reflect your business logic.
| Tab | Description |
| ----------------- | ------------------------------------------------------------------------------------------------------------------------------------- |
| **Total Records** | Custom SQL query that defines what counts as the total record population. Leave empty to use the full row count from Select criteria. |
| **Failed Rows** | Custom SQL query that defines which records are considered failed/invalid. Leave empty to use the rows that fail the Condition rules. |
Both tabs support **Prizm AI suggestions** — click the AI icon to generate an appropriate query based on the metric context.
Choose how Prizm determines when to raise an alert:
**Automated Threshold** — Prizm learns statistical baselines from historical data and alerts when invalid row counts deviate by a configurable number of standard deviations:
| Priority | Deviation range | Default |
| ------------ | --------------- | ---------------------------------- |
| **Low** | 0σ – 1σ | Off — within expected variation |
| **Medium** | 1σ – 2σ | Off — monitor for drift |
| **High** | 2σ – 3σ | **On** — investigation recommended |
| **Critical** | ≥ 3σ | **On** — immediate action required |
**Custom Threshold** — Set explicit upper and/or lower bounds on the invalid row count or score to trigger alerts.
Scroll down past the Threshold section to complete:
* **Semantic context**: Assign Domain, Application, Product, Tag, and Dimension to make the metric discoverable and contextually relevant.
* **Owners**: Set Business owner, Technical owner, and Steward.
* **Audience**: Restrict who can view this metric's results.
Click **Save Changes** when done.