Documentation

The Filter Data operator is a versatile tool in Self-Serve Data Transformation, designed to apply filters to any dataset or transformation result.

How Does It Work?

1. ‍Select the Table or Dataset
   Start by choosing the data you want to filter. You can select any table in your database. Alternatively, you can filter results from previous transformation steps, enabling seamless multi-step workflows.‍
2. ‍Apply Filters
   Add filters to define the subset of data you need: ou can create simple or compound conditions for filtering. You can also combine multiple conditions using AND/OR logic to refine the filtering criteria.

3. ‍Add Table-Specific Filters
   You can apply filters at two levels:‍‍

• ~Dataset Level: Filter the entire dataset from the previous step using global conditions.
• ~Table Level: Add specific filters to individual tables within the dataset by clicking «Add Table Filter» and defining conditions for the selected table.

4. ‍Run and Save
   Once your filters are configured, click «Run and Save» to apply the filters and update the model. The resulting dataset will reflect the specified filters.

Use Case

Filtering Cross-Channel Data with Google Ads Performance Max Campaigns

To illustrate, let’s consider a practical example:

Goal

Prepare Cross-Channel data that includes:‍

• Google Ads data filtered by account.
• Performance Max campaigns filtered by Campaign Type.

Steps‍‍

1. ‍Select the Dataset: Choose the result from the previous step (e.g., Cross-Channel data).‍
2. Apply Dataset-Level Filters: Add a filter to the entire dataset to include only data for a specific account_name = Improvado.
3. ‍Add Table-Level Filters: Click Add Table Filter to apply a filter specifically to the Google Ads table, and add a condition: campaign_advertiser_type = Performance Max.‍
4. Run and Save: Execute the filter operation and update the model.

As a result we have entire dataset is filtered by account name. The Google Ads table includes only Performance Max campaigns. This ensures a clean, non-duplicated dataset, ready for downstream analysis or transformations.

Conclusion

The Filter Data operator brings precision and flexibility to your data transformation workflows. By allowing you to apply filters at both the dataset and table levels, it ensures your data is clean, targeted, and ready for deeper analysis.