Ansicht
Dokumentation

ABAPSHEET_ABAP_SQL_HIERARCHIES - SHEET ABAP SQL HIERARCHIES

RFUMSV00 - Advance Return for Tax on Sales/Purchases   PERFORM Short Reference  
This documentation is copyright by SAP AG.

: Working with Hierarchies

This cheat sheet summarizes the functions, ABAP SQL offers together with ABAP CDS for working with hierarchical data that is stored in database tables. Hierarchical data in database tables means that lines of one or more database tables are connected by parent-child relationships. There are many use cases where hierarchical data plays a role and where accessing information about the hierarchical relationship is important. For example, a common task can be to find out the descendants or ancestors of a given hierarchy node or to aggregate values of subtrees.

Overview

In former times you had to load the data from the database into internal tables and program it all by yourself (if you did not find an appropriate API). In between, meshes offered some features for working with hierarchies, as shown in this example, but have not found wide distribution.

Meanwhile, the standard AS ABAP database is a SAP HANA database that offers a lot of helpful features. Among other things, you will find a set of hierarchy functions there that allow you to deal with hierarchical data directly on the database and that you can look up in the SAP HANA documentation. Now you might expect that you must use AMDP in order to access these functions from your ABAP programs, but no need to do so! ABAP SQL and ABAP CDS support hierarchies directly by wrapping the HANA built-in functions without any loss of performance. You can stay in the comfortable ABAP world and nevertheless have access to most modern features. All you have to do, is to understand some concepts and learn some additional syntax and then you can start right away.

SQL Hierarchies

With SQL hierarchy we denote a special hierarchical data source that you can use in the FROM clause of ABAP SQL queries. A SQL hierarchy is a tabular set of rows which form the hierarchy nodes of a hierarchy and which contains additionally hierarchy columns that contain hierarchy attributes with hierarchy-specific information for each row. For creating a SQL hierarchy, you need the following:

• Data Source

This can be any data source you can access normally in an ABAP SQL query, as most commonly a database table or a CDS view, but also a CTE (common table expression). The structure and content of the data source should be able to represent hierarchical data.

• Parent-Child Relation

A parent-child relation must be defined between two or more columns of the data source. From the parent-child relation and the actual data of the data source, the SQL hierarchy consisting of parent nodes and child nodes can be created. The parent-child relation must be defined by a self-association which we call hierarchy association. This can be achieved with CDS associations or CTE associations. A data source exposing a hierarchy association can be used as a hierarchy source for creating a SQL hierarchy.

• Hierarchy Creation

From a hierarchy source, that is a data source exposing a hierarchy association, a SQL hierarchy can be created. This can be done either by defining a CDS hierarchy outside an ABAP program or with the hierarchy generator of ABAP SQL directly in the FROM clause of an ABAP SQL query inside an ABAP program.

The following topics show you step-by-step how SQL hierarchies can be created and accessed.

Creating SQL Hierarchies

ABAP CDS Hierarchies

With CDS hierarchies, you outsource the hierarchy data source and the creation of the SQL hierarchy from your ABAP program to ABAP CDS. Here the hierarchy is a fully fledged CDS entity, it is reusable in different programs or in other CDS entities (views), and can be part of your data model including access control using CDS DCL. For a CDS hierarchy, the hierarchy source cannot be anything else but a CDS view that exposes a hierarchy association. Here is a very simple example for that:

This CDS view entity accesses the database table DEMO_SIMPLE_TREE, where the actual data reside, and exposes a self-association _tree. The ON condition of the association defines a parent-child relation between the elements id and parent. It simply means that a row of the result set where column parent has the same value as the column id of another row is a child of the latter row in the hierarchy that is constructed from that view. The CDS view exposes also another column name of the database table that represents the remaining data content. Note that you can define such CDS views for any available data source and that the ON condition can be more complex than shown in the simple example here.

Now we can use the above CDS view as the hierarchy source of a CDS hierarchy that can be defined as follows:

The CDS DDL statement DEFINE HIERARCHY that can be used in the DDL source code editor of ADT defines a CDS hierarchy as a CDS entity that can be accessed in CDS views or ABAP SQL as a SQL hierarchy. The most important additions of the statement are:

• SOURCE for specifying the hierarchy source, here our DEMO_CDS_SIMPLE_TREE_VIEW.

• CHILD TO PARENT ASSOCIATION for specifying the hierarchy association, here _tree.

• START WHERE for defining the root nodes of the SQL hierarchy, here represented by an input parameter p_id that must be passed when accessing the CDS hierarchy.

• SIBLINGS ORDER BY to define also a sort order for sibling nodes besides the sort order that comes from the parent-child relationship anyhow.

• An element list { ... } that defines the columns of the SQL hierarchy, here simply all elements of the hierarchy source.

For a full description and all other additions see DEFINE HIERARCHY.

When you access the CDS hierarchy, all lines are selected from the original data source, in our case the database table DEMO_SIMPLE_TREE, that fulfill the START WHERE condition. Those make up the root node set of the SQL hierarchy. In the simplest case we have exactly one root node, but more are possible. Then, for each root node, its descendants are retrieved. That means each line from the database table that fulfills the ON-condition of the hierarchy association is added to the SQL hierarchy. And for each descendant this is done again and again until all descendants are found. And that is basically all! Further additions to DEFINE HIERARCHY allow you to control the creation of the SQL hierarchy, for example, whether multiple parents are allowed or how orphans or cycles should be handled.

Besides the elements of the hierarchy, the element list can also contain the hierarchy attributes listed under Hierarchy Attributes. Then the SQL hierarchy is enriched with columns containing information about the role, the current line plays as a hierarchy node, as, for example, the hierarchy rank or the hierarchy level. In our example, we did not add such elements, because ABAP SQL does that implicitly for you when accessing the CDS hierarchy!

The SQL hierarchy can be used in an ABAP SQL query by using the CDS hierarchy directly as a data source of the FROM clause:

And although we did not define any hierarchy attributes in the element list of the CDS hierarchy, we can add all the hierarchy columns listed under Hierarchy Columns to the SELECT list of our ABAP SQL statement! This is always possible when a SQL hierarchy is accessed in ABAP SQL. We can pass any ID to the CDS hierarchy now and see what happens. If such a line is found in the database table, the respective hierarchical data will be retrieved and delivered. Execute program DEMO_SQL_HIERARCHIES for filling the database table with randomly generated data and inspect the tabular result. As expected, all elements of the SELECT list appear as columns. Note that the content of column NAME could be anything. It is filled here with a string representation of the path from the root node to the current node for demonstration purposes only.

From the ABAP language point of view, CDS hierarchies are the most convenient way of using SQL hierarchies. Now let us turn to other ways, involving more ABAP, until we do not use any CDS more in the end.

ABAP SQL Hierarchy Generator HIERARCHY

The ABAP SQL hierarchy generator is a ABAP SQL function named HIERARCHY, that allows you to define a SQL hierarchy in the ABAP program itself. Let us look directly at an example:

ASSERT asql_cds_result = cds_result.

Looks familiar? Well, almost the same syntax used for defining the CDS hierarchy is used in the brackets HIERARCHY( ... ) and it does exactly the same! The difference is the same as it is between ABAP SQL joins and joins in CDS views:

• If you code it in ABAP SQL, it is for usage in one program only.

• If you code it in ABAP CDS, it is for usage in many programs or whole data models.

And, as you can see, we dare to prove this with an ASSERT statement. Please also note that we use the hierarchy columns again. They exist implicitly when an SQL hierarchy, here created by the hierarchy generator, is accessed.

The above hierarchy generator of ABAP SQL accesses the same hierarchy source as the CDS hierarchy, namely the CDS view DEMO_CDS_SIMPLE_TREE_VIEW that exposes the necessary hierarchy association _tree. In the following code snippet, we replace the CDS hierarchy source with a CTE:

ASSERT asql_cte_result = cds_result.

Common table expressions (CTEs) are a very powerful tool for defining subqueries that can be used in subsequent queries of the same WITH statement. They can be regarded as an internal ABAP SQL definition of data sources that fulfill the same functionality as program external data sources, especially CDS views. As you see above, the CTE cte_simple_tree_source does the same as the CDS view DEMO_CDS_SIMPLE_TREE_VIEW:

• It accesses the database table DEMO_SIMPLE_TREE.

• It exposes an association _tree by using the addition WITH ASSOCIATIONS.

The main query of the WITH statement uses the hierarchy generator in the same way as the SELECT above, just with the CTE as a data source instead of the CDS view and the result is - of course - the same.

For a full description of the hierarchy generator and all other additions see SELECT, FROM HIERARCHY.

We managed to create a SQL hierarchy with ABAP SQL means only. Last but not least we will use CTEs as hierarchies themselves. You might skip the following section and turn directly to the hierarchy navigators if you are not too interested in this syntactic gimmicks.

ABAP CTE Hierarchies

A CTE that produces hierarchical data can declare itself as a SQL hierarchy of a freely defined name with the addition WITH HIERARCHY. That simply means that subsequent queries of the same WITH statement can use the CTE as a hierarchy with its implicit hierarchy columns or - more important - in hierarchy navigators.

The following code snippets show the three ways in which a CTE can produce hierarchical data:

ASSERT cte_cds_result  = cds_result.
ASSERT cte_asql_result = cds_result.
ASSERT cte_cte_result  = cds_result.

A CTE that is exposed as a SQL hierarchy must access a SQL hierarchy itself and in the end these are always based on a CDS hierarchy or the ABAP SQL hierarchy generator as shown above. Again, the hierarchy source of the hierarchy generator can be a CDS view or a CTE exposing the hierarchy association. Running DEMO_SQL_HIERARCHIES shows that all assertions are fulfilled.

Hierarchy Navigators

Hierarchy navigators are an additional set of hierarchy functions in ABAP SQL that allow you to work on existing SQL hierarchies instead of creating them. Hierarchy navigators can work on SQL hierarchies created as shown above, namely on CDS hierarchies, the hierarchy generator or a CTE hierarchy. They can be used as data sources in ABAP SQL queries. If you need a SQL hierarchy multiple times, from a performance point of view it is best to create it once with a given set of root nodes and then access it with hierarchy navigators. Furthermore, each hierarchy navigator can add further hierarchy columns to the result set that offer additional options for the evaluation.

In the following examples, we access our CDS hierarchy with hierarchy navigators. But you could also replace it with the hierarchy generator or a CTE hierarchy. Check the examples of the documentation, where this is also shown.

Hierarchy Node Navigator HIERARCHY_DESCENDANTS

As the name says, HIERARCHY_DESCENDANTS fetches all descendants for any nodes from a SQL hierarchy. It adds HIERARCHY_DISTANCE as an additional hierarchy column to the result set. Let us look at an example. All examples are code snippets from DEMO_SQL_HIERARCHIES again.

Our CDS hierarchy DEMO_CDS_SIMPLE_TREE_VIEW is used to create a SQL hierarchy with a start node passed to parameter p_id and for a node sub_id all descendants are fetched. Running the program shows the result including the additional column HIERARCHY_DISTANCE that contains the distance to the respective start node. A further parameter DISTANCE - not shown here - allows you to restrict the distance to the respective start node.

Hierarchy Node Navigator HIERARCHY_ANCESTORS

Now the other way around: ABAP SQL function HIERARCHY_ANCESTORS returns the ancestors of any given node of an existing hierarchy:

Looking at the result when running DEMO_SQL_HIERARCHIES, you see that the value of column HIERARCHY_DISTANCE is negative now. Using aggregate functions or evaluating the internal result table, you can now easily extract further information like the number of ancestors and so on.

Hierarchy Node Navigator HIERARCHY_SIBLINGS

Besides descendants and ancestors, hierarchy nodes also can have siblings, that is nodes that have the same parent node. You can find these by looking for all nodes with the same value in hierarchy column HIERARCHY_PARENT_RANK. But there is also HIERARCHY_SIBLINGS as a hierarchy function for that:

You see that this function adds another hierarchy column HIERARCHY_SIBLING_DISTANCE that contains the distance to the respective start node. Running DEMO_SQL_HIERARCHIES, where we start with a node that definitely has some siblings, shows the result.

Hierarchy Aggregate Navigators

Finally let us turn to the hierarchy aggregate navigators that allow you to apply some aggregate functions to descendants and ancestors of any node of a SQL hierarchy:

• HIERARCHY_DESCENDANTS_AGGREGATE

• HIERARCHY_ANCESTORS_AGGREGATE

We will show an example for the descendants case and refer to the documentation for the ancestors.

Applying aggregate functions to columns normally means that you have some data there for which this makes sense. In our simplistic SQL hierarchy tree we do not have such meaningful data. On the other hand, this can also be a use case: You can have the administrative data for the parent-child relation in one database table and the real data in another one. And for that use case, the hierarchy aggregate navigator HIERARCHY_DESCENDANTS_AGGREGATE gives you the option of joining such data to your SQL hierarchy:

In our example, we join an internal table value_tab of the same program to the SQL hierarchy. In a real life example you would join another database table, of course. On the other hand the example shows ABAP SQL's capability of using internal tables as data sources. You even can go so far to evaluate hierarchical data in internal tables with ABAP SQL by using an internal table as data source for a CTE hierarchy!

The example does the following:

• We use the hierarchy aggregate navigator HIERARCHY_DESCENDANTS_AGGREGATE as a data source of a FROM clause.

• Our CDS hierarchy DEMO_CDS_SIMPLE_TREE_VIEW joined with internal table value_tab is used as the data source.

• The ABAP SQL function returns a tabular result of nodes of the data source.

• The aggregate function SUM behind MEASURES sums up the values of the column amounts of the joined internal table for all descendants of each node returned by the ABAP SQL function.

• The WHERE condition restricts the result set by a freely programmable condition.

• The WITH additions add further rows to the result set that can be recognized by values in an additional hierarchy column HIERARCHY_AGGREGATE_TYPE:

• WITH SUBTOTAL

In the row where HIERARCHY_AGGREGATE_TYPE has value 1, column AMOUNT_SUM contains the sum of the values of all hierarchy nodes that meet the WHERE condition.

• WITH BALANCE

In the row where HIERARCHY_AGGREGATE_TYPE has value 2, column AMOUNT_SUM contains the sum of the values of all hierarchy nodes that do not meet the WHERE condition.

For more WITH additions see the documentation.

Running DEMO_SQL_HIERARCHIES shows the result. It also shows the result of the joined data source, where you can check that the calculated values are correct.

Working with Hierarchies

For the complete reference documentation about SQL hierarchies, see SELECT, FROM hierarchy_data.

Fill RESBD Structure from EBP Component Structure   BAL_S_LOG - Application Log: Log header data  
This documentation is copyright by SAP AG.

Length: 24214 Date: 20240509 Time: 053758     sap01-206 ( 369 ms )