Dataset settings determine how the dataset uses the source data.
Connecting multiple tables
If multiple tables are available in the data source, you can merge them using the JOIN operator.
When the tables are joined, a link is created between them. When you create a link, you specify the fields from the source table and merged table.
You can use the following types of JOIN operators:
Data source operation modes
You can run your dataset in one of two modes, depending on the data source.
Datasets can work with data sources in the following modes:
- Direct access. DataLens executes all the data queries on the data source side.
- Materialization. DataLens pre-saves the data to the database and then works with the data.
In direct access mode, DataLens materializes only part of the data for dataset preview.
All data requests are executed on the data source side.
You can only use direct access for the following data sources:
- Yandex.Metrica API
- AppMetrica API
Materialization is the process of uploading data from the data source to the DataLens database.
You can run materialization as a one-time process or periodically on a schedule.
Data is uploaded to the materialization DB once. Then DataLens only makes queries to the materialized data.
To sync DataLens storage with the source, you can reload the data.
If you use a CSV file as the data source, DataLens automatically materializes the dataset.
Data is uploaded to the DataLens materialization DB according to a schedule. The schedule is set in the dataset settings.
For example, every second week on Thursday at 03:00.
The minimum available interval is once a day.
You can configure permissions for the whole dataset. For more information, see Managing dataset access.
You can also set data permissions at the row level (Row-level security or RLS). Learn more in Managing access to data rows in a dataset.