Custom Metrics

Add and delete custom metrics for tables

PUT /v1/tables/{id}/customMetric

PUT /v1/tables/{id}/customMetric

Python

from metadata.sdk import configure
from metadata.sdk.entities import Tables

configure(
    host="https://your-company.open-metadata.org/api",
    jwt_token="your-jwt-token"
)

table_id = "a1b2c3d4-e5f6-7890-abcd-ef1234567890"

# Add a table-level custom metric
Tables.add_custom_metric(table_id, {
    "name": "active_customer_ratio",
    "description": "Ratio of active customers to total customers",
    "expression": "SELECT CAST(SUM(CASE WHEN active = true THEN 1 ELSE 0 END) AS FLOAT) / COUNT(*) FROM {table}"
})

# Add a column-level custom metric
Tables.add_custom_metric(table_id, {
    "name": "email_domain_count",
    "description": "Number of distinct email domains",
    "columnName": "email",
    "expression": "SELECT COUNT(DISTINCT SPLIT_PART(email, '@', 2)) FROM {table}"
})

# Delete a custom metric
Tables.delete_custom_metric(table_id, "active_customer_ratio")

Response (Add Custom Metric)

{
  "id": "455e3d9d-dbbf-455e-b3be-7191daa825f3",
  "name": "agent_performance_summary",
  "fullyQualifiedName": "sample_data.ecommerce_db.shopify.agent_performance_summary",
  "customMetrics": [
    {
      "name": "high_performer_ratio",
      "description": "Ratio of high-performing agents (score > 90) to total agents",
      "expression": "SELECT CAST(SUM(CASE WHEN performance_score > 90 THEN 1 ELSE 0 END) AS FLOAT) / COUNT(*) FROM {table}"
    }
  ]
}

Custom Metrics

Define custom SQL-based metrics on a table or specific column. Custom metrics are evaluated during profiling runs and tracked over time.

Add Custom Metric

PUT /v1/tables/{id}/customMetric

id · string · required

UUID of the table.

name · string · required

Name of the custom metric. Must be unique within the table.

expression · string · required

SQL expression to evaluate for this metric. Must return a single numeric value.

columnName · string

Name of the column this metric applies to. If omitted, the metric is table-level.

description · string

Description of what this metric measures.

Delete Custom Metric

DELETE /v1/tables/{id}/customMetric/{metricName}

id · string · required

UUID of the table.

metricName · string · required

Name of the custom metric to delete.

PUT /v1/tables/{id}/customMetric

Python

from metadata.sdk import configure
from metadata.sdk.entities import Tables

configure(
    host="https://your-company.open-metadata.org/api",
    jwt_token="your-jwt-token"
)

table_id = "a1b2c3d4-e5f6-7890-abcd-ef1234567890"

# Add a table-level custom metric
Tables.add_custom_metric(table_id, {
    "name": "active_customer_ratio",
    "description": "Ratio of active customers to total customers",
    "expression": "SELECT CAST(SUM(CASE WHEN active = true THEN 1 ELSE 0 END) AS FLOAT) / COUNT(*) FROM {table}"
})

# Add a column-level custom metric
Tables.add_custom_metric(table_id, {
    "name": "email_domain_count",
    "description": "Number of distinct email domains",
    "columnName": "email",
    "expression": "SELECT COUNT(DISTINCT SPLIT_PART(email, '@', 2)) FROM {table}"
})

# Delete a custom metric
Tables.delete_custom_metric(table_id, "active_customer_ratio")

Response (Add Custom Metric)

{
  "id": "455e3d9d-dbbf-455e-b3be-7191daa825f3",
  "name": "agent_performance_summary",
  "fullyQualifiedName": "sample_data.ecommerce_db.shopify.agent_performance_summary",
  "customMetrics": [
    {
      "name": "high_performer_ratio",
      "description": "Ratio of high-performing agents (score > 90) to total agents",
      "expression": "SELECT CAST(SUM(CASE WHEN performance_score > 90 THEN 1 ELSE 0 END) AS FLOAT) / COUNT(*) FROM {table}"
    }
  ]
}

---

Returns

Add returns the updated table object with the custom metric included.

Delete returns no content (204).

Response

customMetrics · array

Array of custom metrics defined on the table.

Show properties

name · string

Name of the custom metric.

description · string

Description of the metric.

expression · string

SQL expression for the metric.

columnName · string

Column name if this is a column-level metric.

---

Error Handling

Code	Error Type	Description
401	UNAUTHORIZED	Invalid or missing authentication token
403	FORBIDDEN	User lacks permission
404	NOT_FOUND	Table or custom metric does not exist
400	BAD_REQUEST	Invalid metric definition or SQL expression
409	CONFLICT	Custom metric with same name already exists