hard_deletes

Available from dbt v1.9 or with Versionless dbt Cloud.

snapshots/schema.yml

snapshots:
  - name: <snapshot_name>
    config:
      hard_deletes: 'ignore' | 'invalidate' | 'new_record'

dbt_project.yml

snapshots:
  <resource-path>:
    +hard_deletes: "ignore" | "invalidate" | "new_record"

snapshots/<filename>.sql

{{
    config(
        unique_key='id',
        strategy='timestamp',
        updated_at='updated_at',
        hard_deletes='ignore' | 'invalidate' | 'new_record'
    )
}}

Description

The hard_deletes config gives you more control on how to handle deleted rows from the source. Supported options are ignore (default), invalidate (replaces the legacy invalidate_hard_deletes=true), and new_record.

 When to use the hard_deletes and invalidate_hard_deletes config?

Use invalidate_hard_deletes (v1.8 and earlier) if:

• You want to invalidate deleted rows by setting their dbt_valid_to timestamp to the current time (implicit delete).
• You are working with smaller datasets where tracking deletions as a separate state is unnecessary.
• Gaps in the snapshot history (missing records for deleted rows) are acceptable.

Use hard_deletes: new_record (v1.9 and higher) if:

• You want to explicitly track deletions by adding new rows with a dbt_is_deleted column (explicit delete).
• You want to maintain continuous snapshot history without gaps.
• You are working with larger datasets where explicitly tracking deleted records improves data lineage clarity.

warning

If you're updating an existing snapshot to use the hard_deletes config, dbt will not handle migrations automatically. We recommend either only using these settings for net-new snapshots, or arranging an update of pre-existing tables before enabling this setting.

Default

By default, if you don’t specify hard_deletes, it'll automatically default to ignore. Deleted rows will not be tracked and their dbt_valid_to column remains NULL.

The hard_deletes config has three methods:

Methods	Description
ignore (default)	No action for deleted records.
invalidate	Behaves the same as the existing invalidate_hard_deletes=true, where deleted records are invalidated by setting dbt_valid_to to current time. This method replaces the invalidate_hard_deletes config to give you more control on how to handle deleted rows from the source.
new_record	Tracks deleted records as new rows using the dbt_is_deleted meta field when records are deleted.

Considerations

• Backward compatibility: The invalidate_hard_deletes config is still supported for existing snapshots but can't be used alongside hard_deletes.
• New snapshots: For new snapshots, we recommend using hard_deletes instead of invalidate_hard_deletes.
• Migration: If you switch an existing snapshot to use hard_deletes without migrating your data, you may encounter inconsistent or incorrect results, such as a mix of old and new data formats.

Example

snapshots/schema.yml

snapshots:
  - name: my_snapshot
    config:
      hard_deletes: new_record  # options are: 'ignore', 'invalidate', or 'new_record'
      strategy: timestamp
      updated_at: updated_at
    columns:
      - name: dbt_valid_from
        description: Timestamp when the record became valid.
      - name: dbt_valid_to
        description: Timestamp when the record stopped being valid.
      - name: dbt_is_deleted
        description: Indicates whether the record was deleted.

The resulting snapshot table contains the hard_deletes: new_record configuration. If a record is deleted and later restored, the resulting snapshot table might look like this:

id	dbt_scd_id	Status	dbt_updated_at	dbt_valid_from	dbt_valid_to	dbt_is_deleted
1	60a1f1dbdf899a4dd...	pending	2024-10-02 ...	2024-05-19...	2024-05-20 ...	False
1	b1885d098f8bcff51...	pending	2024-10-02 ...	2024-05-20 ...	2024-06-03 ...	True
1	b1885d098f8bcff53...	shipped	2024-10-02 ...	2024-06-03 ...		False
2	b1885d098f8bcff55...	active	2024-10-02 ...	2024-05-19 ...		False

In this example, the dbt_is_deleted column is set to True when the record is deleted. When the record is restored, the dbt_is_deleted column is set to False.