dbt source snapshot-freshness has been renamed to
dbt source freshness. Its node selection logic is now consistent with other tasks. In order to check freshness for a specific source, you must prefix it with
- Snowflake: Turn off transactions and turn on autocommit by default. Explicitly specify
commit for DML statements in incremental and snapshot materializations. Note that this may affect user-space code that depends on transactions.
manifest.json uses a
v3 schema that includes additional node properties (no changes to existing properties)
run_results.json uses a
v3 schema that has added
skipped as a potential
sources.json has a new
v2 schema that has added timing and thread details
New and changed documentation
source: Updated selection logic to match other tasks. When selecting a specific source to check freshness, you must prefix it with
- Node selection syntax, commands: Switch
--select across the board. (Commands which previously used the
--models flag still support it for backwards compatibility.)
- YAML selectors now support an optional
default property. If set, dbt will use custom selection criteria for commands that do not specify their own selection/exclusion flags.
- Selection methods and state comparison caveats: Add
state:modified subselectors, and reflect that it now includes changes to upstream macros.
- Test selection examples includes more discussion of indirect selection (a change in v0.20), and the optional "greedy" flag/property (new in v0.21), which you can optionally set to include tests that have a mix of selected + unselected parents
Elsewhere in Core
- Postgres profile property
connect_timeout now configurable. Also applicable to child plugins (e.g.
- Redshift: profile property
ra3: true to support cross-database source definitions and read-only querying
- BigQuery: profile property
execution_project now configurable. Snapshots support
target_dataset config aliases.