Column Content Conditions

A Column Content condition matches the dominant data format detected in a column’s sampled values (a sampled subset; see Sampling), useful for distinguishing columns that hold structured documents, archives, images, or encoded payloads from columns that hold plain values, such as classifying a column only when it contains JSON, PDF, or PNG content.

Fields

• Content format(s) – content_format (a single format) or content_formats (a list to match any of several). See the supported formats below.
• Match Threshold – the share of the column’s non-null sampled values a single dominant format must account for, shown as a percentage in the UI (defaults to 80%) and stored as a fraction from 0.0-1.0 in the API/JSON (dominance_threshold, default 0.8). The condition passes only when that dominant format is also one of the selected formats. Because the threshold tests for a dominant format, a column split evenly between two formats may not pass at the default threshold even if one of them is selected.

Supported formats:

• Text / structured: JSON, XML, HTML, YAML, CSV, Base64, Plain text
• Documents: PDF, DOCX, XLSX, PPTX, Legacy MS Office, RTF
• Archives: ZIP, GZIP
• Images: PNG, JPEG, GIF, TIFF
• Columnar: Parquet

Example

This condition matches a column whose sampled values are dominantly JSON or XML:

{ "target": "CONTENT_TYPE", "content_formats": ["JSON", "XML"], "dominance_threshold": 0.8 }

Behavior

Format detection runs on the raw sampled values, so binary signatures are preserved. Empty and all-null columns are handled without error.

In Match Confidence scoring, a Column Content match is outranked by Google DLP, Snowflake Native, Amazon Comprehend, Regex, and Data Entropy matches, and outranks a Column Name, Data Location, Data Length, or Column Size match on the same column.

Availability

Column Content evaluates on Snowflake ALTR Native jobs using ALTR Hosted processing, on OLTP jobs, and on Databricks jobs with a collection attached (the ALTR Native option). It doesn’t evaluate on Snowflake In-Warehouse jobs. Content detection requires ALTR Hosted processing, the same constraint as Google DLP. See Classification Processing Location for the full picture.