Skip to main content

unique_key

dbt_project.yml
snapshots:
<resource-path>:
+unique_key: column_name_or_expression

Description

A column name or expression that is unique for the inputs of a snapshot. dbt uses this to match records between a result set and an existing snapshot, so that changes can be captured correctly.

In Versionless and dbt v1.9 and later, snapshots are defined and configured in YAML files within your snapshots/ directory. The unique_key is specified within the config block of your snapshot YAML file.

caution

Providing a non-unique key will result in unexpected snapshot results. dbt will not test the uniqueness of this key, consider testing the source data to ensure that this key is indeed unique.

Default

This is a required parameter. No default is provided.

Examples

Use an id column as a unique key

You can also specify configurations in your dbt_project.yml file if multiple snapshots share the same unique_key:

dbt_project.yml
snapshots:
<resource-path>:
+unique_key: id

Use a combination of two columns as a unique key

This configuration accepts a valid column expression. As such, you can concatenate two columns together as a unique key if required. It's a good idea to use a separator (e.g. '-') to ensure uniqueness.

Though, it's probably a better idea to construct this column in your query and use that as the unique_key:

0