Summarize groups without depending on row adjacency

A report needs a count per carrier. LOOP AT ... GROUP BY forms groups by a key even if matching rows are not adjacent in the original table. You can bind group-key values and, when needed, loop through the members.

Bind a key and the group size

LOOP AT dispatches INTO DATA(dispatch)
  GROUP BY ( carrier = dispatch-carrier size = GROUP SIZE )
  INTO DATA(carrier_group).
  DATA(result) = carrier_group-size.
ENDLOOP.

Each group iteration exposes one carrier_group; GROUP SIZE supplies its member count. To process each original row within the group, add LOOP AT GROUP carrier_group INTO DATA(member). If you only need the group key and aggregate metadata, omit member processing.

Grouping syntax has several variants and release constraints. This original example demonstrates the idea; consult the target release documentation before adapting it. For simpler totals or one row-by-row action, LOOP AT, REDUCE, or ABAP SQL aggregation may be easier to review.

Sources

• SAP Samples: Internal Tables — grouping
• ABAP Keyword Documentation: LOOP AT GROUP BY

TYPES: BEGIN OF ty_dispatch,
         dispatch_id TYPE string,
         carrier TYPE string,
       END OF ty_dispatch.
TYPES ty_dispatches TYPE STANDARD TABLE OF ty_dispatch WITH EMPTY KEY.
DATA(dispatches) = VALUE ty_dispatches( ( dispatch_id = 'D-1701' carrier = 'NORTH' ) ( dispatch_id = 'D-1702' carrier = 'NORTH' ) ( dispatch_id = 'D-1703' carrier = 'EAST' ) ).
LOOP AT dispatches INTO DATA(dispatch)
  GROUP BY ( carrier = dispatch-carrier size = GROUP SIZE )
  INTO DATA(carrier_group).
  DATA(result) = carrier_group-size.
ENDLOOP.

WRITE result.

out->write( result ).