AT - Control break with extracts

Variants

1. AT NEW f.
2. AT END OF f.
3. AT FIRST.
4. AT LAST.
5. AT fg.

Effect

In a LOOP which processes a dataset created with EXTRACT , you can use special control structures for control break processing. All these structures begin with AT and end with ENDAT . The sequence of statements which lies between them is then executed if a control break occurs.

You can use these key words for control break processing with extract datasets only if the active LOOP statement is proceesing an extract dataset.

The control level structure with extract datasets is dynamic. It corresponds exactly to the sort key of the extract dataset, i.e. to the order of fields in the field group HEADER by which the extract dataset was sorted .

At the end of a control group ( AT END OF , AT LAST ), there are two types of control level information between AT and ENDAT :

If the sort key of the extract dataset contains a non-numeric field h (particularly in the field group HEADER ), the field CNT(h) contains the number of control breaks in the (subordinate) control level h .

For extracted number fields g (see also ABAP/4 number types ), the fields SUM(g) contain the relevant control totals.

Notes

The fields CNT(h) and SUM(g) can only be addressed after they have been sorted. Otherwise, a runtime error may occur.
The fields CNT(h) and SUM(g) are filled with the relevant values for a control level at the end of each control group ( AT END OF , AT LAST ), not at the beginning ( AT FIRST , AT NEW ).
When calculating totals with SUM(g) , the system automatically chooses the maximum field sizes so that an overflow occurs only if the absolute value area limits are exceeded.
You can also use special control break control structures with LOOP s on internal tables.

Variant 1

AT NEW f.

Variant 2

AT END OF f.

Effect

f is a field from the field group HEADER . The enclosed sequence of statements is executed if

the field f occurs in the sort key of the extract dataset (and thus also in the field group HEADER ) and

the field f or a superior sort criterion has a different value in the current LOOP line than in the prceding ( AT NEW ) or subsequent ( AT END OF ) record of the extract dataset.

Example


DATA: NAME(30),
      SALES TYPE I.
FIELD-GROUPS: HEADER, INFOS.
INSERT: NAME  INTO HEADER,
        SALES INTO INFOS.
...
LOOP.
  AT NEW NAME.
    NEW-PAGE.
  ENDAT.
  ...
  AT END OF NAME.
    WRITE: / NAME, SUM(SALES).
  ENDAT.
ENDLOOP.

Notes

If the extract dataset is not sorted before processing with LOOP , no control level structure is defined and the statements following AT NEW or AT END OF are not executed.
Fields which stand at hex zero are ignored by the control break check with AT NEW or AT END OF . This corresponds to the behavior of the SORT statement, which always places unoccupied fields (i.e. fields which stand at hex zero) before all occupied fields when sorting extract datasets, regardless of whether the sort sequence is in ascending or descending order.

Variant 3

AT FIRST.

Variant 4

AT LAST.

Effect

Executes the relevant series of statements just once - either on the first loop pass (with AT FIRST ) or on the last loop pass (with AT LAST ).

Variant 5

AT fg.

Addition

... WITH fg1

Effect

This statement makes single record processing dependent on the type of extracted record.

The sequence of statements following AT fg are executed whenever the current LOOP record is created with EXTRACT fg (in other words: when the current record is a fg record).

Addition

... WITH fg1

Effect

Executes the sequence of statements belonging to AT fg WITH fg1 only if the record of the field group fg in the dataset is immediately followed by a record of the field group fg1 .

Index
� SAP AG 1996