New columns in table function output are no longer treated as behavior changes¶

Beginning September 4, 2026, new columns in the output of the built-in table functions in the INFORMATION_SCHEMA, ACCOUNT_USAGE, ORGANIZATION_USAGE, and READER_ACCOUNT_USAGE schemas are no longer treated as behavior changes.

This change applies only to the built-in table functions in these schemas. It doesn’t apply to other built-in table functions, such as FLATTEN, RESULT_SCAN, and SPLIT_TO_TABLE.

When a new column is added to the output of a built-in table function, the change will no longer be included in a behavior change bundle, and the change will no longer be announced with the other behavior changes. These types of changes will be introduced in server releases and feature updates.

If you depend on an exact column list, introducing a new column may cause unexpected problems. In this scenario, consider one of these approaches:

• As a workaround, you can temporarily exclude that column from the output of the table function. This capability is available now, so you can set up column overrides before this change takes effect on September 4, 2026.
• As a long term solution, update your scripts to select specific columns from the table function output.