Skip to main content

updated_at

dbt_project.yml
snapshots:
  <resource-path>:
    +strategy: timestamp
    +updated_at: column_name

Description​

A column within the results of your snapshot query that represents when the record row was last updated.

This parameter is required if using the timestamp strategy. The updated_at field may support ISO date strings and unix epoch integers, depending on the data platform you use.

Default​

No default is provided.

Examples​

Use a column name updated_at​

Coalesce two columns to create a reliable updated_at column​

Consider a data source that only has an updated_at column filled in when a record is updated (so a null value indicates that the record hasn't been updated after it was created).

Since the updated_at configuration only takes a column name, rather than an expression, you should update your snapshot query to include the coalesced column.

Using updated_at with check strategy​

When using the check strategy, dbt verifies the rows in the source table based on a timestamp column to determine freshness. By default, it uses the updated_at column if it's explicitly configured. If the updated_at column isn't configured, dbt falls back to using the current timestamp for all rows.

Check out the following example showing how to configure the updated_at column for the check strategy:

sources:
  - name: my_source
    tables:
      - name: my_table
        freshness:
          warn_after: {count: 12, period: hour}
          error_after: {count: 24, period: hour}
          strategy: check
          updated_at: updated_at_column_name

In this example:

  • strategy: check tells dbt to use the updated_at column for verifying row freshness.
  • updated_at: updated_at_column_name explicitly sets the column (pdated_at_column_name) to be used for the freshness checks. If the updated_at_column_name column isn't present in the source table, dbt will fall back to the current timestamp for all rows.
0
Edit this page