Identifying similar columns

During sensitivity scans and schema change scans, Tonic Structural identifies groups of similar columns.

To identify similar columns, Structural uses a text embedding model to calculate the semantic similarity between any two column names in the database. When a column name's semantic similarity to the name of a given column is above a specified threshold, then the column is similar to the given column.

If a column has similar columns, then the Applied Generator column contains an icon that includes the count of similar columns.

By default, the similar columns icon is hidden. To display the similar columns icon, hover over the column row.

When you assign a generator to a column, the similar columns icon for that column remains visible during your current session.

When you click the similar columns icon, Structural displays a panel with an option to filter the list to display the current column and its similar columns. To apply the filter, click Filter.

The similar columns filter is applied, and other column filters are removed. Table filters remain in place.