Overview of multi-column rules

Multi-column rules allow you to validate relationships between two or more columns within a table — ensuring consistency, uniqueness, or logic across combined values. Depending on the data format, different rule types are available.

That's how yo create a multi-column rule:

Numeric (float) rules

These rules apply to numeric column combinations.

Available rule types

• Equals

• Does not equal

• Greater than

• Greater than or equal

• Less than

• Less than or equal

• Unique

• Custom SQL

• Custom Python

Example: Equals condition

Check whether one numeric column equals a combination of others:

“Awarded Amount” must equal “Consumed Amount” + “Remaining Amount”

Example: Unique combination

Ensure that the combination of three numeric values is unique.

Example: Custom SQL

Write a SQL query to apply custom multi-column validation.

String-based rules

String rules allow comparisons between columns or uniqueness checks.

Two-column comparisons

• Contains

• Starts with

• Ends with

Check if “Department” starts with “Department Code”.

Multi-column string rules

• Unique

• Custom SQL

• Custom Python

Each combination of “Contract Type” + “Contract Title” must be unique.

Example: Custom Python

Write pandas logic to check formatting, duplication, or mapping rules.

Date-based rules

Used to check sequences or consistency across timestamp or date columns.

Two-column comparisons

• After

• Before

• After or equal

• Before or equal

Ensures logical time progression between two date fields

Multi-column date rules

• Unique

• Custom SQL

• Custom Python

Prevent duplicate entries with the same start and end dates

Example: SQL-based date validation

Find gaps or overlaps in booking periods or schedule windows

Notes

• SQL and Python rules return a row per failed match

• You can add filters to scope the rule to specific rows or conditions

• Multi-column logic is useful for validating derived fields, time sequences, or complex integrity checks