Google DLP Conditions

A Google DLP condition matches infoType detections that Google’s Sensitive Data Protection service (formerly Cloud DLP) finds in a column’s sampled values, joining Amazon Comprehend and Snowflake Native as the classifier rule builder’s external, ML-based signals. Google DLP ships with roughly 150 built-in infoTypes covering categories such as financial identifiers, government IDs, healthcare data, and credentials, and can distinguish a sensitive value from similar-looking data using surrounding context — unlike a Regex Condition, which matches a fixed pattern.

Fields

• infoType(s) — a searchable multi-select of Google DLP infoTypes. Selecting multiple infoTypes combines them with OR logic — the condition passes if any selected infoType is detected at or above the Likelihood Threshold.
• Likelihood Threshold — the minimum likelihood a detection must meet, from VERY_UNLIKELY to VERY_LIKELY. Defaults to LIKELY.

Note

Each column needs at least 10 non-null, non-empty values to be evaluated. Columns with fewer values are skipped.

How Google DLP Accesses Your Data

Which Google Cloud account runs the detection depends on where the classification job processes:

• On Snowflake and Databricks, under ALTR Hosted processing, ALTR operates the Google Cloud account and credentials used for detection — no Google Cloud setup is required on your side.
• On OLTP, via the OLTP Classification Agent, you supply your own Google Cloud service account credentials and project ID, and detection runs against your own Google Cloud quota and billing.

Sampling and Data De-referencing

Before any data is sent to Google, ALTR applies data de-referencing, using the same column-only sampling every classification job uses: values are drawn from individual columns only, never full rows, and scrambled so no actual row can be reconstructed from what’s sent. Only this de-referenced sample — never the original data — transits to the Google DLP service.

Example

This condition matches columns where Google DLP detects an email address at or above LIKELY:

{ "target": "GDLP", "info_type": "EMAIL_ADDRESS", "minimum_likelihood": "LIKELY" }

Behavior

Google DLP evaluates each column’s de-referenced values against its infoType library and returns a likelihood score, from VERY_UNLIKELY to VERY_LIKELY, for each detection. ALTR includes only detections that meet or exceed the condition’s Likelihood Threshold.

Of all condition types, a Google DLP match ranks highest in Match Confidence scoring.

Availability

Google DLP evaluates on Snowflake and Databricks ALTR Native jobs using ALTR Hosted processing — not In-Warehouse processing — and on OLTP ALTR Native jobs. On processing paths where it can’t run, the condition is treated as not evaluated rather than failing the job. See Classification Processing Location for the full picture.

On OLTP, Google DLP runs through your own Google Cloud account rather than ALTR’s. See the OLTP Classification Agent for how ALTR handles credentials for this and other external conditions.