Implementing Variables

Objectives
After completing this lesson, you will be able to:

After completing this lesson, you will be able to:

  • Implement variables to filter data by attributes
  • Map variables and input parameters

Variables

You define a variable in a calculation view when you want to pass dynamic values to filter attribute columns.

When a calculation view that contains a variable is called by a front-end reporting tool, the calculation view passes a request to the front-end tool to provide a value (usually using a user prompt) to complete a WHERE clause. The WHERE clause is added to the query that runs on top of the calculation view. For example, if the user was prompted to choose a country and the user chose 'JP', the SQL that runs on the calculation view looks like this:

SELECT <columns> FROM <calculation_view_name> WHERE (("COUNTRY" = 'JP')) GROUP BY <group_by columns>;

Creating Variables

You define variables in the Semantics node of a Calculation View, in the Parameters tab.

The View/Table Value Help setting is used to define which table/view SAP HANA will fetch the data to show in the prompt. The Reference Column defines which (unique) column in this table/view will provide the possible values for the variable.

By default, the View/Table Value Help setting refers to the current calculation view. This means that the values of the specified attribute column, for example CUSTOMER (customer ID), will be scanned in order to propose the various values in the prompt.

A different approach is to define a dedicated source table/view for the variable. For example, you could define the master data table CUSTOMERS as the Value Help source for the variable, and - of course - choose the column from this master data table that matches the column of your calculation view you want to filter on: in our example, the Customer ID.

When the Value Help table and one column are chosen, you assign the APPLY FILTER attribute. That is, you identify which attribute column will be filtered at runtime, based on the variable values chosen by the user.

Note
When the Value Help source table/view is the calculation view itself, the reference column you define is automatically assigned to the APPLY FILTER section as a filter attribute. However, you can define another attribute if needed, provided that it is consistent with the reference column.

The behavior of variables at runtime depends on whether an entry is required for the variable or not. These scenarios can be defined as follows:

  • If a variable is defined as mandatory, the user needs to provide the values, ranges, or intervals at runtime.

  • For non-mandatory variables, if nothing is specified at runtime, all the data for the corresponding attributes is returned by the view without filtering.

Creating Variables (2)

Creating Variables (3)

More than one value can be chosen for a variable when you select the Multiple Entries checkbox.

Note

In the Data Preview, From and To are displayed in the Variable Values dialog even when a variable has not been defined as range.

Value Help for Variables

When a dialog box appears to the user they must make a selection. However, rather than an empty field appearing and the user having to guess valid values or figure out the format for a value (for example, is the country code UK or GB?), we can have a dialog box populated with the run-time values available. To do this, you make a selection in the setting View/Table for value help. The default entry is the calculation view where the variable is being created. This means that you present all possible values from the column that is assigned to the Attribute setting in the variable definition.

While this might seem like a great idea, remember that the list may be huge and would be difficult for the user to navigate. Imagine presenting a list that includes every employee in a very large organization? If the list should be restricted to offer limited values, such as employees in your department, then you should reference an external calculation view or a table that exposes a restricted list. It is also good practice, from a performance perspective, to refer to a restricted list from an external calculation view or table.

Value Help Based on Hierarchies

When creating a variable on an attribute that is associated with one or several hierarchies, you can specify one of the hierarchies in the variable definition. With this option, the user can navigate the hierarchy, rather than a flat list, to select the values in the value help. This makes navigation much easier when there are a lot of values; imagine being able to first select your country, then your department, before the list of employees appears? You can use either parent-child or level hierarchies.

The following are some of the basic rules that apply when implementing value help based on hierarchies:

  • If you refer to a parent-child hierarchy, the variable attribute column must be defined as the parent attribute and not the child.

  • If you refer to a level hierarchy, the variable attribute column must be defined at the leaf level; that is, the bottom level.

Note
Variables are not relevant for measures.

Use Variables

Mapping of Variables and Input Parameters

Pushing Down Input Parameters and Variables to Lower Level Calculation Views

In many cases, calculation views use other calculation views as data sources. This is not necessarily confined to two levels; we can go on and layer the calculation views to create a stacked model. When you execute a calculation view in which variables or input parameters are defined, it is possible to pass their values (entered by the end user at runtime) to the lower level calculation views. In fact, the input parameters and variables at the lower levels are usually ignored unless you define input parameters and variables at the top level and map them to the input parameters and variables in the lower levels. This is called parameter mapping and is an important feature of SAP HANA calculation view modeling.

To enable parameter mapping, you must use the Input Parameter/Variables Mapping feature. You can find this feature in the Parameters tab in the calculation view.

When you open the mapping pane, you must first select the type of mapping you want to work with using the dropdown selector.

Parameter Mapping Types

There are four types of parameter mapping and you choose the type from the Manage Mapping pane.

When you make your type selection, you will then see (on the left side) the input parameters and variables that are defined in the calculation views from all lower layers in the stack, which are related to the mapping type you selected. On the right side, you will see the input parameters and variables that are defined in the current calculation view (the one you are editing).

Note
You can only map Variables to Variables and Input Parameters to Input Parameters. Cross-Mapping (such as an Input Parameter to a Variable) is not possible.

You simply drag a line between the left and right side to map them. There is also an auto-map feature which means that if the names are the same, the mapping is done with a single click. The auto-map feature generates the input parameters or variables for the current view with the same name as the source variables and also maps them. This means that you do not have to manually create the input parameters or variables in the current view.

Pushing filters down to the source views using parameter mapping is a common scenario. To enable this, choose the type Data Sources from the dropdown list in the Manage Mapping dialog.

Hint
Mapping parameters of the current view to the parameters of the underlying data sources moves the filters down to the underlying data sources during runtime, which reduces the amount of data transferred across them. This is a great way to improve performance.

Another common scenario is when you want to push parameters down from the main calculation view to a calculated column in a lower view to support a calculation. Again, this would be the type Data Sources.

Note that in the type of mapping Data Source, you only map input parameters to input parameters. In other words, a Variable defined in an underlying Calculation View cannot be mapped to a Variable defined in the current Calculation View. However, it is possible to access these variables from the Extract Semantics feature and copy them to the current view. To do that, you right-click the data source in the calculation scenario and choose Extract Semantics. Then choose the Variables tab and select the ones you want to copy to the semantics of your Calculation View.

Mapping Input Parameters to SQL View Parameters

If a data source for your calculation view is an SQL view and the SQL view requires parameters, you can provide the values from input parameters of the consuming calculation view.

This technique is often used to push down the filters down to the SQL view in order to reduce the workload of the Calculation View.

Note
Mapping Input Parameters from SQL views is possible from SAP HANA QRC 4/2021 QRC onwards.

Mapping Input Parameters to Remote Views (Smart Data Access)

From SAP HANA Cloud QRC 1/2022 onwards, it is possible to pass input parameters from a Calculation View to an SQL View or a Calculation View that is accessed remotely as a Virtual Table. This ensures that filters are pushed down and executed as close to the data sources as possible.

This new feature requires Smart Data Access (SDA), relying on the odbc adapter – it is not possible with remote sources accessed through the Smart Data Integration (SDI) adapter. The scenario supports remote SAP HANA databases running on SAP HANA Cloud, or SAP HANA On-Premise version 2.0 minimum.

Note
Besides, in SAP HANA On-Premise, Input Parameters in SQL Views are supported only from version 2.0 SPS02 onwards.

Mapping for External Value Help Views

Another important use case for mapping input parameters and variables is to enable dynamic value help views.

When you define input parameters and variables, the default data source that generates the value help list is taken from the calculation view itself. So, essentially you are getting an unrestricted list of all possible values to choose from. However, you can also redirect the value help to use a list from another table or view. The main reason we do this is to expose a restricted value help list.

This is also good practice for performance because the value help is not competing with the main calculation view for data. For example, you could create a calculation view on a table that contains all possible cities. Here, your calculation view would include a fixed filter expression that restricts the cities to a specific country. This means that the value help list presents only cities of a specific country to the user.

What if you wanted to change the country? You could go back to the calculation view and change the fixed filter expression, but this would be inefficient.

Cascading Prompts Architecture

What we should do is replace the fixed value in the filter expression with a variable based on country. Then, we should map this variable to a variable we define in the main calculation view for the country. This means that when a user is prompted for a country in the main view, the value chosen is passed through the mapping to the value help calculation view, so that the cities are filtered by the country that was chosen. The list of cities is then presented as the value help for the cities column. This is also known as cascading prompts. Cascading is not restricted to two levels; you can also cascade prompts across multiple levels. For example, you could prompt for Continent, which then restricts the list of Countries, which in turn restricts the list of Cities, and so on.

To implement value help parameter mapping, you must select the option Views for value help for variables/input parameters from the dropdown list in the Manage Mapping dialog.

Note
An external view based on a hierarchy could also be considered as a value help cascading solution, and might be more visually appealing to the user.

Log in to track your progress & complete quizzes