Work Areas in Open SQL Statements

The Open SQL statements DELETE, INSERT, MODIFY, UPDATE, and SELECT * (without the CORRESPONDING FIELDS addition), as well as FETCH (if the database cursor was opened with the preceding SELECT statement), work with work areas wa or internal tables itab. The following prequisites apply in these statements to the work area or row type of the internal table; if the prerequisites are ignored, a syntax error or an exception will occur.

The first two points are minimum prerequisites that were replaced as of Release 6.10 in Unicode programs or in the treatment of strings in database tables through stricter requirements:

• The data type of the work area or row type of the internal table must not be deep or contain deep components. This excludes strings as of Release 6.10.

• The work area or rows of the internal table must be at least as long as the database structure and the alignment must match. For the DELETE statement, this prerequisite must only be met in the length of the primary key.

• If the work area or rows of the internal table are structured, the Unicode fragment view of the database structure must match that of the work area or internal table in Unicode programs.

• If the work area or rows of the internal table are elementary, they must be character-type and flat in Unicode programs. The columns of the database structure must also be character-type and flat.

• If the work area contains strings as components or is itself a string, or if columns of type SSTRING, STRING, or RAWSTRING

Note

The work area or row structure of the internal table should always be built like the database structure. If you work in the Open SQL statement with a single database table or one view, a data object built the same way can be declared with reference to the relevant structure in the ABAP Dictionary. If you work with more than one database table (in the SELECT statement), a data object built the same way can be built as nested structure, which contains, as substructures, the structures of the single database tables or views in the sequence in which they are listed in the statement. No structure should be used in which all components lie on one level, as possible alignment gaps between the single database tables or views are not taken into account.