Dbt-clickhouse

Bug Fixes
* An [issue](https://github.com/ClickHouse/dbt-clickhouse/issues/348) was detected when using multiple projections. We solved it and added a test to cover that use case. ([#349](https://github.com/ClickHouse/dbt-clickhouse/pull/349))

Documentation
* A [CONTRIBUTING.md](CONTRIBUTING.md) file was added to the repo. Please follow the instructions prior contributing.

New Features
* [ClickHouse projections](https://clickhouse.com/docs/en/sql-reference/statements/alter/projection) are now fully supported for `table` materialization, and partly supported for `distributed_table` materialization.
The projection config should be added to the model config ([342](https://github.com/ClickHouse/dbt-clickhouse/pull/342)), for instance:
python
{{ config(
materialized='%s',
projections=[
{
'name': 'your_projection_name',
'query': 'your_projection_query'
}
]
) }}


Bug Fixes
* Until this release, when writing tests, it was not possible to pass empty seed data. ([341](https://github.com/ClickHouse/dbt-clickhouse/pull/341))
* When a cluster was used, the adapter left a few `__dbt_backup` tables in the schema. After the fix, these backup tables are now properly dropped. ([326](https://github.com/ClickHouse/dbt-clickhouse/pull/326))
* Due to [this GitHub change](https://github.com/actions/runner-images/issues/9692), we needed to adjust the `docker compose` command in our tests. ([#334](https://github.com/ClickHouse/dbt-clickhouse/pull/334))

Improvements

* Added a new **experimental** incremental strategy, `insert_overwrite`, which replaces existing data in a target table partition by partition. This ensures that only the specified partitions are overwritten, helping to maintain performance and data consistency. Note that this feature has not been tested with distributed models and may not work with such materializations.
[Anton Bryzgalov](https://github.com/bryzgaloff) Thank you for your contribution! ([#201](https://github.com/ClickHouse/dbt-clickhouse/pull/201))
* Schema changes for incremental models were improved and now include a `sync_all_column` option.
[Can Bekleyici](https://github.com/canbekley) huge thanks for your contribution. ([#332](https://github.com/ClickHouse/dbt-clickhouse/pull/332))
* Previously, view settings were only applied during view creation, not when querying the view. Starting with this release, the settings attribute will now also be applied when running queries on a view. ([324](https://github.com/ClickHouse/dbt-clickhouse/pull/324))
* Enhanced the clickhouse__listagg macro to support single field ordering with optional direction, ensuring compatibility with ClickHouse's sorting limitations. Added validation to prevent the use of multiple order-by fields. ([318](https://github.com/ClickHouse/dbt-clickhouse/pull/318))

Documentation
* Update the docs with the new `insert_overwrite` incremental strategy. ([331](https://github.com/ClickHouse/dbt-clickhouse/pull/331))
* Add documentation for codec column configuration. ([317](https://github.com/ClickHouse/dbt-clickhouse/pull/317))

Bug Fix
* Refresh materialized_view table only if `--full-refresh` is specified.
* Fix temporary table creation to support dbt unit tests.
* ClickHouse dictionaries are now correctly created "on cluster" when a cluster is defined.
Improvements
* Added database prefix option for local tables of distributed tables (until this change, only a suffix was supported).
* You can now customize tcp_keepalive configuration for native connections.
* Implement listagg ([groupArray](https://clickhouse.com/docs/en/sql-reference/aggregate-functions/reference/grouparray)) macro.

Improvements
- Upgrade the connector to use dbt-core 1.8.0. More info about this upgrade can be found [here](https://github.com/dbt-labs/dbt-adapters/discussions/87).

Beginning in v1.8, dbt-core and adapters are decoupled. Going forward, your installations should explicitly include both dbt-core and the desired adapter. The new pip installation command should look like this:

pip install dbt-core dbt-clickhouse

Bug Fix
- Fix bool_or behavior (a cross-database dbt macro ).
- Fix query_settings for models with contracts.
- Fix the option to use Nullable and LowCardinality in column constraints.
- Specifying an empty or null value for a connection_overrides field produces invalid DDL for the dictionary materialization. A fix was introduced in this release.

Improvements
- The connector is now supporting [column codecs](https://clickhouse.com/blog/optimize-clickhouse-codecs-compression-schema#specialized-codecs).

Bug Fix
- A bug in (experimental) Distributed Table model creation could lead to errors when there was a change in the model definition (see, e.g.,
https://github.com/ClickHouse/dbt-clickhouse/issues/226). Thanks to [Thomas Schmidt](https://github.com/Somtom) for the Fix!
- A comment at the end of a model would break the query used to retrieve the result column datatypes. Thanks to [triou](https://github.com/tevariou)
for the bug report and the fix. Closes https://github.com/ClickHouse/dbt-clickhouse/issues/256

Improvements
- The new materialization for ClickHouse dictionaries now takes an optional "credentials dictionary" argument that overrides the
global credentials values for user, password, database, host, and port (including removing any of those values by adding empty values if not needed).
This allows better control over creating dictionaries on different server. Thanks to [Cristhian Garcia](https://github.com/Ian2012)
for the PR!
- A new `ttl` setting has been added to model configuration that will insert the provided ClickHouse TTL expression in the appropriate place.
Thanks to [Evan Rusackas](https://github.com/rusackas) for the contribution!
- The Agate library should now be lazy loaded. This should modestly improve dbt startup times (after dbt-clickhouse is upgraded to dbt 1.8.x).
Thanks to [Daniel Reeves](https://github.com/dwreeves) for PR.