Currently, we only support dataset mapping functionality. Other reporting validation functions will be added in the future.
In AML, the Modeling and Reporting Layers are vital for data modeling and reporting. Changes in the Modeling Layer can impact report functionality and accuracy.
To address this, you can use Reporting Validation to locate and validate reporting items linked to models and datasets, ensuring reference accuracy during and after changes in your project.
When to use Reporting Validation
- Fix errors after deployment: Identify and fix the broken reporting items caused by changes in the Modeling Layer.
- Find and replace names of fields, models, or datasets: If you want to change the naming convention of your fields, models, or datasets, Reporting Validation can help you locate all instances where they are referenced and replace them with the new names.
How to use Reporting Validation
Run Reporting ValidationTo ensure the integrity of your Reporting Layer and identify potential issues or compromises, follow these steps:
- Step 1: Navigate to the Reporting Validation tab, located in the bottom panel.
- Step 2: Click the
Validate Reporting Itemsbutton. This action will generate a list that highlights Reports, Dashboard Filters, Data Alerts, and other components that may be affected by changes in the Modeling Layer.
When using Reporting Validation, you'll encounter three types of Validation Results:
- Error: This indicates that issues exist in the references, even before any changes are made within the Modeling Layer.
- Warning: A warning signifies that dependencies might break when modifications are made and deployed to the Reporting Layer. It serves as an early alert for possible issues.
- Fixed: Any prior errors made will be corrected when you deploy your changes to Production.
Addressing Issues with Broken Items
After executing Reporting Validation, if there are existing or potential issues related to broken reporting items, we propose two possible solutions:
Option 1: Modify the object names in the Modeling layer based on errors identified from the Reporting Validation results:
- Correct the names of your dataset, model, or field (dimension/measure) objects.
- This supports continued interaction between the Reporting Items and the reference name in the Modeling layer.
Option 2: Replace fields post-project deployment if option 1 is not desirable:
- After deploying the project, navigate to the problematic reporting items.
- Edit these items by replacing the broken fields with valid fields.
- Save your changes once completed.
Please note: This approach may cause some downtime for your reporting items. It's crucial to inform your Dashboard users beforehand.
If any widget (report and filter) in reporting references deleted/renamed datasets, the error will be raised and you will need to fix the error before continuing with the deployment.
Changes that might cause this error:
- Alter the dataset name in Data Modeling
- (AML 1.0) Remove one of the existing datasets in the
- (AML 2.0) Remove one of the existing datasets in the Modeling 4.0.
For example, I change my dataset name from
ecommerce_khai in my dataset file and index.aml file.
Deploy to production, the Error will be raised because there are reports created from that dataset in Reporting.
You will need to either Fix the error or Cancel the Deployment to continue.
By clicking on Update, you will be able to update your widgets and point them to another dataset listed in the file
Best Practices When Using Reporting Validation
- Use Reporting Validation before and after you make any changes in the Modeling Layer. This will help you prevent errors from happening and fix them quickly if they do.
- Carefully review validation results, focusing on Errors and Warnings indicating issues, and check Fixed results for corrections.
- Use Reporting Validation to ensure consistent naming conventions by finding and replacing names in all related reporting items.
Does Holistics provide error validation at field and model levels?
Holistics currently does not provide error validation at the field and model levels.