POST_START

Classifying Sensitive Columns with Tags for Attribute-Driven Governance

I recently took on the responsibility of improving data governance within our organization. One of the key areas I focused on was ensuring that sensitive data, such as personal information, was properly classified and tagged. This is essential for compliance, security, and data lifecycle management.

Identifying Sensitive Columns

I started by reviewing the structure of our customer master table in the production database. This table contains a variety of customer attributes, some of which are clearly sensitive. I needed to identify which columns should be tagged with a data classification label to ensure they are handled appropriately.

I ran the DESCRIBE TABLE EXTENDED command to get a detailed view of the table schema and column descriptions. This helped me understand the context of each column and determine which ones might contain personally identifiable information (PII).

DESCRIBE TABLE EXTENDED production.customers.customer_master;

<

col_name	data_type	comment
customer_id	bigint	customer identifier
customer_name	string	customer display name
region	string	sales region

I noticed that the customer_name and region columns are descriptive, but the email and phone columns are not included in the schema output. This was a clue that those columns might be part of the table but were not described in the metadata. I decided to proceed with tagging those columns as they are commonly used to store sensitive information.

Tagging Sensitive Columns with Data Classification

I began by tagging the email column with the data_classification tag set to pii. This is a standard practice for identifying personally identifiable information in the data catalog.

ALTER TABLE production.customers.customer_master ALTER COLUMN email SET TAGS ('data_classification' = 'pii');

I verified the change by checking the catalog metadata to ensure the tag was applied correctly. I wanted to make sure that the tag was visible and that the system recognized the classification. This step is crucial for downstream tools and processes that rely on metadata to enforce access controls and data handling policies.

Tagging Additional Sensitive Columns

Next, I moved on to the phone column, which is also typically used to store personal contact information. I applied the same data_classification tag to this column as well.

ALTER TABLE production.customers.customer_master ALTER COLUMN phone SET TAGS ('data_classification' = 'pii');

By tagging these columns, I was able to create a clear and consistent way to identify sensitive data within the customer master table. This classification helps ensure that these columns are handled according to the appropriate security and compliance policies.

Conclusion

Through this process, I learned the importance of metadata tagging in data governance. By classifying sensitive columns with tags, I enabled better visibility, control, and compliance across the organization. This is just one part of a broader strategy for attribute-driven governance, but it’s a critical step in ensuring that our data is managed responsibly and securely.