DataWorks:Data warehouse layering

Built-in layers

By default, DataWorks provides five layers: Operational Data Store (ODS), Dimension (DIM), Data Warehouse Detail (DWD), Data Warehouse Summary (DWS), and Application Data Service (ADS).

The functions of each layer are as follows:

Custom layers

Go to the Data Warehouse Layering page to view the preset data warehouse layers. Switch between Tiled Display and Hierarchy Display to adjust the page layout.

If the preset layers do not meet your needs, create custom layers:

Configure data warehouse layer checkers

Data warehouse layer checkers enforce naming conventions for models and metrics. When you create a model or metric, the checker can automatically generate an object name, making the physical table name easy to understand.

1. Click the target layer to go to its details page.

2. In the Checker Management section, click Create Rule and select Create Rule for Model or Create Rule for Metric.

3. The checker parameters are described in the following tables.

   Create Rule for Model

   Configuration item	Description
   Rule Name	The name of the new checker rule.
   Rule Type	Defaults to Name. When you create a model in Dimensional Modeling, you can select the corresponding checker to automatically generate a table name.
   Rule Definition	Define the naming convention for the data model. This rule is used to check the model names within the layer during Dimensional Modeling. Enter a prefix for the table to check against. For example, if the data warehouse layer requires table names to start with dim, you can enter dim. You can also click the icon to define the rule using elements such as business category abbreviation, data domain abbreviation, data mart abbreviation, or subject area abbreviation. You can combine these elements freely. The following are recommended naming conventions: DIM: dim_{business_category}_{data_domain}_{custom_content}_{storage_policy} DWD: dwd_{business_category}_{data_domain}_{business_process}_{custom_content}_{storage_policy} DWS: dws_{business_category}_{data_domain}_{custom_content}_{time_period} ADS: ads_{business_category}_{data_mart}_{subject_area}_{custom_content}_{time_period}
   Description	The description of the checker.

   Create Rule for Metric

   Configuration item	Description
   Rule Name	The name of the new checker rule.
   Rule Type	Configurable for Display Name or Abbreviation. When you create a metric, the system generates a metric name based on the default rule.
   Rule Definition	Define metric names by selecting a modifier, atomic metric, and statistical period in any order. In Data Metric, this rule checks metric names within the layer.
   Description	A description of the checker.

4. Set the default checker.

   In Default Rule, set a default checker for models or metrics. If no rule is specified when you create a model or metric, the default checker automatically generates names.

5. Weak and strong rules.

   1. Weak Rule: If a naming checker is set as a weak rule, you can bypass the rule and still save the model with a custom table name.

   2. Strong Rule: If a naming checker is set as a strong rule, you must adhere to the naming convention when creating a model or metric. Otherwise, you cannot save the model or metric.

Use checkers

1. Use in Dimensional Modeling or Data Metric.

   1. Dimensional Modeling

      1. Auto-generate a table name: When creating a logical model, select a checker in Naming Rule. The system suggests a name based on the rule.

      2. Validate table names: When creating or editing a table name, if you do not select a checker, the system checks whether the table name conforms to the checker rules when you save the logical model.

      3. If multiple checkers exist, validation passes if the name conforms to any one of them.

   2. Data Metric

      1. Auto-generate a metric name: When creating a derived metric, select Period, Modifiers, and Atomic Metric, then click Intelligent Recommendation. The system will automatically generate metric names based on the default rule.

      2. Validate table names: When creating or editing the Abbreviation and Display Name of a metric and saving it, the system automatically checks if the entered names conform to the checker rules.

      3. If there are multiple checkers, the validation passes as long as the name conforms to any one of them.

2. Start Check and View Results.

   After configuring a checker, click Start Check to verify whether the data model names in the current layer conform to the naming conventions. You can then address any non-compliant models based on the results.

   Note

   • The Start Check function is not currently supported for metric rules.

   • Checks run only for enabled checkers. New checkers are enabled by default. If you do not need to use a checker, you can disable it to prevent it from running checks.

   After the check is complete, click View Results to review the check results.