SAP LOOP GROUP BY ABEXA

Get Example source ABAP code based on a different SAP table

ABAP_ITAB - Grouping Internal Tables, Step by Step
The example is a step-by-step introduction to grouping with LOOP AT GROUP BY.

ABAP_SOURCE_CODE
DEMO CL_DEMO_LOOP_AT_ITAB_GROUP_BY

ABAP_DESCRIPTION
Similar to SQL's GROUP BY, there is also a GROUP BY for working with internal tables that can be used behind LOOP AT itab or in the form IN GROUP in a table iteration with FOR. It replaces the clumsy group level processing with statements AT NEW ... that relies on the order of table columns and content that is sorted respectively.
While there are more examples for all the powerful features of GROUP BY and IN GROUP under

Examples of Grouping with LOOP,

Examples of Grouping with FOR,
the above example shows the grouping of internal tables step by step using a very simple case of an internal table spfli_tab that is filled with data from the database table SPFLI. The following steps explain how the content of the internal table can be grouped using LOOP AT GROUP BY:

Grouping by One Column
The simplest form of grouping is by one column without explicitly specifying the output behavior of the group loop: LOOP AT spfli_tab INTO wa
GROUP BY wa-carrid.
... wa-carrid ...
ENDLOOP.
Within the loop, there is access to the work area wa, in particular to the component wa-carrid that is used for grouping. The work area wa contains the first line of each group and represents the group in the loop. This is called representative binding.
To access the members of a group, a member loop can be inserted into the group loop: LOOP AT spfli_tab INTO wa
GROUP BY wa-carrid.
...
LOOP AT GROUP wa INTO DATA(member).
... member-... ...
ENDLOOP.
...
ENDLOOP.
The member loop is executed using the group represented by wa and its members are assigned to member and are available in the member loop.

Grouping by More than One Column
To group by more than just one criterion, a structured group key is defined as follows. In the simplest case, the grouping criteria are columns of the internal table: LOOP AT spfli_tab INTO wa
GROUP BY ( key1 = wa-carrid key2 = wa-airpfrom ).
... wa-carrid ... wa-airpfrom ...
ENDLOOP.
This is also a representative binding in which the work area wa is reused in the group loop to access the group key.
To access the members of the groups, the exact same member loop can be inserted as when grouping by one column.

Group Key Binding when Grouping by One Column
By explicitly specifying an output area for the group key, a group key binding can be defined explicitly instead of the representative binding in which the output area of the group loop is reused: LOOP AT spfli_tab INTO wa
GROUP BY wa-carrid
INTO DATA(key).
... key ...
ENDLOOP.
The difference to the example with representative binding is the INTO addition after GROUP BY. Instead of reusing wa , an elementary data object key represents the group. This can be generated inline. The additions GROUP SIZE, GROUP • , and WITHOUT MEMBERS can only be used in the group key binding, which gives it more functions than the representative binding. If these are not required, the representative binding can be used. The group key binding can also be used to make the use of the group key in the loop more explicit.
Inserting a member loop works in the same way as in the representative binding, with the difference that a group is now addressed by key instead of wa. LOOP AT spfli_tab INTO wa
GROUP BY wa-carrid
INTO key.
...
LOOP AT GROUP key INTO member.
... members ...
ENDLOOP.
...
ENDLOOP.

Group Key Binding when Grouping by More than One Column
Finally, the group key binding for structured group keys: LOOP AT spfli_tab INTO wa
GROUP BY ( key1 = wa-carrid key2 = wa-airpfrom )
INTO DATA(key).
... key-key1 ... key-key2 ...
ENDLOOP.
Here, key is a structure with the components key1 and key2. A member loop can be inserted in exactly the same way as when grouping by one column.
If the group members are not relevant, the addition NO MEMBERS can be used to save time and memory. LOOP AT spfli_tab INTO wa
GROUP BY ( key1 = wa-carrid key2 = wa-airpfrom
index = GROUP • size = GROUP SIZE )
WITHOUT MEMBERS
INTO DATA(key).
... key-key1 ... key-key2 ... key-index ... key-size ...
ENDLOOP.
It is no longer possible to use a member loop here. Instead, the group key was enriched with optional components for further information using GROUP • GROUP SIZE.