Classifiers & Collections

ALTR Native classification identifies sensitive data by running a collection (a reusable group of classifiers) against a data source. The same collection works across every processing location classification can run in, including the OLTP Classification Agent and ALTR Hosted or In-Warehouse processing on Snowflake.

• A classifier identifies one specific type of data, such as Social Security numbers or email addresses, using one or more classifier rules combined with AND/OR logic. Each rule holds one or more conditions, which can be nested and individually negated with NOT.
• A collection is a named group of one or more classifiers that you attach to a classification job. A collection must contain at least one classifier, and a classifier can belong to multiple collections.

ALTR also provides an ALTR Managed collection: a set of classifiers ALTR creates and maintains for common personally identifiable information (PII), protected health information (PHI), and Payment Card Industry (PCI) data types, and expands over time. Import it to start classifying immediately, or use it as a base to build your own collections from.

Classification Settings

Each classifier carries settings that control how it evaluates sampled data. Both settings apply only to condition types that evaluate sampled row data: Regex, Data Length, Column Content, Google DLP, and Amazon Comprehend conditions. They don’t apply to Column Name, Data Location, or Column Size conditions, which evaluate column metadata rather than sampled data.

• Minimum Match Threshold — the minimum percentage of sampled data that must match for a column to appear in the classification report. The default is 75%; ALTR Managed classifiers default to 50%.
• Sample size — a classifier’s default sample size, set either as a fraction of the table (table %, default 10%) or a fixed number of rows (default 1,000).

If a classification report doesn’t produce the results you expect, adjust these settings and classify the data again.

Create a Classifier

To create a classifier:

1. Select Data Classification in the navigation menu.
2. Click Create New to create a classifier from the All Classifiers page, or open a collection under Collections and click Add Classifier > Create New to create it within that collection. Either way, the classifier is also added to the full classifier list so it can be reused in another collection.
3. Enter a Classifier Name. The name must be unique and cannot begin with ALTR. That prefix is reserved for ALTR Managed classifiers.
4. Add one or more classifier rules and choose whether the classifier matches when any or all of them are true.
5. (Optional) Enter a Description of the type of data the classifier detects.
6. Adjust the classification settings if the defaults don’t fit.
7. (Optional) Click + Add to Collection to add the classifier to a collection.
8. Click Save. ALTR saves the classifier and adds it to the All Classifiers list.

Edit a Classifier

Editing a custom classifier updates its rules, description, classification settings, or collection assignments. ALTR Managed classifiers can’t be edited; to rename a custom classifier, delete it and create another with the new name.

To edit a classifier:

1. Locate the classifier in the All Classifiers list or within a collection.
2. Click the classifier’s ellipsis menu.
3. Select Edit Classifier. ALTR displays a dialog with the classifier’s settings.
4. Update the classifier.
5. Click Save. ALTR saves your changes.

Delete a Classifier

Deleting a custom classifier removes it from every collection it belongs to and from ALTR. ALTR Managed classifiers can’t be deleted.

To delete a classifier:

1. Locate the classifier in the All Classifiers list or within a collection.
2. Click the classifier’s ellipsis menu.
3. Select Delete Classifier. A confirmation dialog displays.
4. Click Delete Classifier to confirm. ALTR removes the classifier from the page immediately, though the classifier count elsewhere in the UI can take several minutes to update.

To delete more than one classifier at a time, select them in the All Classifiers list, click Delete, and confirm.

Collections

Collections group classifiers by data type, compliance requirement, or business use case — for example, one collection for PII classifiers, another for classifiers aligned to PCI requirements.

Create a Collection

To create a collection:

1. Select Data Classification in the navigation menu.
2. Click Create a New Collection.
3. Enter a Collection Name. The name must be unique and cannot begin with ALTR.
4. (Optional) Enter a Description of the type of data the collection covers.
5. Click Next.
6. (Optional) Select classifiers to add to the collection.
7. Click Next, review the collection’s details, then click Save. ALTR creates the collection.

A collection’s name can’t be changed after it’s created; to rename one, duplicate it and save the copy with a new name.

Add Classifiers to a Collection

To add a single classifier: open the collection under Collections, click Add Classifier > Add Existing, select the classifier, and click Save. ALTR adds the classifier to the collection.

To add classifiers in bulk: select them in the All Classifiers list, click Add to Collection, select the collection, and click Save. ALTR adds the selected classifiers to the collection.

To remove a classifier from a collection without deleting it from ALTR: open the collection, click the classifier’s ellipsis menu, and select Remove from Collection. ALTR removes the classifier from the collection.

Tip

When searching for a classifier to add, start with ALTR Managed - to find ALTR Managed classifiers.

Import the ALTR Managed Collection

Once imported, ALTR Managed classifiers can be added to any of your own collections, and the collection can be duplicated and modified to fit your needs. The collection itself can’t be deleted, and its classifiers can’t be edited or deleted.

To import the ALTR Managed collection:

1. Select Data Classification in the navigation menu.
2. Click ALTR Managed under Collections.
3. Click Import ALTR Managed Collection. ALTR imports the collection into your account.

As ALTR adds classifiers to the ALTR Managed collection over time, sync your copy to pick up the additions.

To sync the ALTR Managed collection:

1. Select Data Classification in the navigation menu.
2. Click ALTR Managed under Collections.
3. Click Sync Classifiers. If the button is disabled, the collection is already up to date; otherwise ALTR opens a dialog showing the classifiers to sync.
4. (Optional) Click View Latest Classifiers to see which classifiers were added since the last sync.
5. Click Sync. ALTR adds the new classifiers to your copy of the collection.

Note

If you generated a classification report using the ALTR Managed collection before syncing, reclassify the data afterward so the report includes the newly added classifiers.

Duplicate a Collection

Duplicating a collection creates a copy, including all of its classifiers. This is useful for customizing a set of classifiers for a different data source, testing changes without affecting the original, or creating variations for different teams.

To duplicate a collection:

1. Select Data Classification in the navigation menu.
2. Click the ellipsis menu for the collection under Collections.
3. Select Duplicate Collection. ALTR displays a dialog to configure the copy.
4. Enter a Collection Name and, optionally, a Description, then click Next.
5. Review the collection, then click Save. ALTR creates the duplicate collection.

Delete a Collection

Deleting a collection doesn’t delete its classifiers; they remain available to other collections. The ALTR Managed collection can’t be deleted once imported.

To delete a collection:

1. Select Data Classification in the navigation menu.
2. Click the ellipsis menu for the collection under Collections.
3. Select Delete Collection. A confirmation dialog displays.
4. Click Delete Collection to confirm. ALTR deletes the collection.

Running a Collection

Attach a collection to a classification job to run it. ALTR samples the data source, evaluates each classifier’s rules against the sample, and includes a column in the classification report when its match threshold is met.

Where the job executes depends on the data source:

• OLTP databases (Oracle, SQL Server, PostgreSQL, MySQL) run on the OLTP Classification Agent in your own environment.
• Snowflake runs at a processing location you choose. See Classification Processing Location.
• Databricks runs on ALTR Hosted classification. See Classification Processing Location.

Sampling

Sampling is random and drawn from individual columns only, never from full rows. The sampled data is scrambled and can’t be reconstructed or used to identify a specific record.

Match Confidence

The classification report shows a Match Confidence badge (High, Medium, or Low) for every match. Hovering the badge shows a one-sentence explanation of which signals contributed to the score, including any rules that were evaluated but excluded from it.

From highest to lowest, Match Confidence ranks: Google DLP, Snowflake Native, Amazon Comprehend, Regex, Column Name, Data Location, and Data Length or Column Size (tied lowest).