More detailed error messages for sensitive data classification

Use cases

Imagine you run automatic classification jobs to scan sensitive data across your tables and some runs fail. With this update, you can pinpoint the exact cause of the failure without guessing.

• Query standardized error messages for classification failures.
• Identify original Snowflake error codes appended to error strings.
• Streamline troubleshooting for failing tables or views.

How to use

SELECT table_name, status, error_message, last_classification_attempt
FROM snowflake.account_usage.data_classification_latest
WHERE error_message IS NOT NULL;

What happens in this code:

1. Queries the account usage view to check recent sensitive data classification attempts.
2. Filters specifically for tables or objects that encountered failures.
3. Returns the table name, status, detailed error message, and associated Snowflake error code.

What it returns

Column	Description
TABLE_ID	Internal system-generated identifier for the table.
TABLE_NAME	Name of the table.
SCHEMA_ID	Internal system-generated identifier for the schema.
SCHEMA_NAME	Name of the schema.
DATABASE_ID	Internal system-generated identifier for the database.
DATABASE_NAME	Name of the database.
RESULT	Latest classification result as a JSON object.
STATUS	Status of classification, either CLASSIFIED or REVIEWED.
TRIGGER_TYPE	Mode of the trigger (MANUAL or AUTO CLASSIFICATION).
LAST_CLASSIFIED_ON	Timestamp of the last successful classification.
LAST_CLASSIFICATION_ATTEMPT	Timestamp of the last classification attempt.
ERROR_MESSAGE	Descriptive error message from the last failed attempt including the originating error code.

Limitations and requirements

• Latency for this view might be up to three hours.
• The view retains data for as long as the underlying table exists.