英文原文：SAP-samples · 上游版本 d43eb13c4c7c · 原始 Markdown

Structures

• Structures
  • Introduction
  • Globally Available Structures and Structured Types
  • Creating Structures and Structured Types Locally
    • Creating Structured Types
    • Creating Structures
    • Creating Structures Using Existing Structured Types
    • Creating Structures by Inline Declaration
    • Creating Constant and Immutable Structures
    • Creating Enumerated Structures
    • Creating Anonymous Structures
  • Variants of Structures
  • Accessing (Components of) Structures
    • ASSIGN Statements
  • Populating Structures
    • Using the VALUE Operator
    • Using the NEW Operator
    • Using the CORRESPONDING Operator and MOVE-CORRESPONDING Statements
  • Clearing Structures
  • Processing Structures
    • Structures in ABAP SQL Statements
    • Structures in Statements for Processing Internal Tables
  • Including Structures
  • Generic Structured Types
    • Fully Generic Structured Types
    • User-Defined Partially Generic Structured Types
  • Excursions
    • sy Structure
    • Getting Structured Type Information and Creating Structures at Runtime
    • Boxed Components
    • Recursive Structure References
  • Executable Example

Introduction

Structures ...

• are data objects with structured data types (which is a complex data type because it is composed of other data types).
• consist of a sequence of components of any data type, that is, the components of a structure can be, for example, elementary data objects, structures themselves, internal tables or references.
• are used to combine different data objects that belong together. A typical example is an address. It has several components, such as name, street, city, and so on, that belong together.
• play an important role in the context of internal tables and database tables. Structured types serve as line types for these tables. Most internal tables across ABAP programs may have structured line types. For database tables, there is no alternative to structured line types.
• can be created locally in an ABAP program and globally. This cheat sheet focuses on locally defined structures and structured types.

⬆️ back to top

Globally Available Structures and Structured Types

• Apart from the local declaration of a structured type, you can create such a type, for example, as global DDIC structure in the ABAP Dictionary. Such a DDIC structure defines a globally available structured type (DDIC type).
• There are other structured types available globally, which may be the structured types most commonly used in ABAP programs:
  • Database tables defined in the ABAP Dictionary can be used as data types just like DDIC structures in an ABAP program. This means that when you create a structure in your ABAP program, for example, you can simply use the name of a database table to address the line type of the table. The structure you created will then have the same structured type as the database table. Typically, you use the database tables to create structures of such a type, or internal tables of such a structured line type, to process data read from the database table in structures or internal tables.
  • Various CDS entities are globally available structured types. For example, a CDS view entity represents a structured data type and can be used as such in ABAP programs (but not in the ABAP Dictionary).
  • Structures and structured data types can be defined in the public visibility section of global classes or in global interfaces and then used globally.

"Creating structures based on globally available structured types
"Database table
DATA struc_from_dbtab TYPE zdemo_abap_fli.
"CDS view entity
DATA struc_from_cds_ve TYPE zdemo_abap_fli_ve.
"CDS abstract entity
DATA struc_from_cds_abs TYPE zdemo_abap_abstract_ent.
"CDS table function
DATA struc_from_cds_tab_func TYPE zdemo_abap_table_function.

"Globally available structured type in the public visibility section of
"classes/interfaces
DATA struc_from_struc_type_in_cl TYPE zcl_demo_abap_amdp=>fli_struc.

"Creating structured types based on globally available structured types
TYPES ty_struc_from_dbtab TYPE zdemo_abap_fli.
TYPES ty_struc_from_cds_ve TYPE zdemo_abap_fli.

[!NOTE]

• This cheat sheet focuses on locally defined structures and structured types.
• Classic DDIC views are not available in ABAP Cloud. They can only be used as structured types in classic ABAP.

⬆️ back to top

Creating Structures and Structured Types Locally

The typical language elements for creating structures and structured types locally in an ABAP program are BEGIN OF ... END OF .... They are used in combination with the TYPES keyword to create a structured type and the DATA keyword to create a structure.

⬆️ back to top

Creating Structured Types

• The following statement defines a structured type introduced by TYPES. The type name is preceded by BEGIN OF (which marks the beginning of the structured type definition) and END OF (the end of the definition).
• The components - at least one must be defined - are listed in between.
• Such structured type definitions are usually grouped together in a chained statement, i.e. TYPES is followed by a colon, and the components are separated by commas.

TYPES: BEGIN OF struc_type,
         comp1 TYPE ...,
         comp2 TYPE ...,
         comp3 TYPE ...,
         ...,
       END OF struc_type.

Alternatively, you can also use the following syntax. However, a chained statement may provide better readability.

TYPES BEGIN OF struc_type.
  TYPES comp1 TYPE ... .
  TYPES comp2 TYPE ... .
  TYPES comp3 TYPE ... .
  ... .
TYPES END OF struc_type.

• The simplest structures and structured types have elementary components.
• As mentioned previously, the components can be of any type, i.e. they can be of structured types themselves, internal table types, or reference types.
• You can use the TYPE and LIKE additions for the types of the components. You can use the LINE OF addition to refer to a table type or an internal table.

TYPES: BEGIN OF struc_type,
         comp1 TYPE i,                  "elementary type           
         comp2 TYPE c LENGTH 5,         "elementary type
         comp3 TYPE structured_type,    "structured type
         comp4 TYPE itab_type,          "internal table type
         comp5 TYPE ddic_type,          "DDIC type
         comp6 TYPE REF TO i,           "data reference
         comp7 LIKE data_object,        "deriving type from a data object                  
         comp8 TYPE LINE OF itab_type,  "component has structured type, type derived from internal table type 
         comp9 LIKE LINE OF itab,       "component has structured type, type derived from internal table
         comp10 TYPE REF TO struc_type, "recursive structure refererence        
         ...,
       END OF struc_type.

[!NOTE]

• Outside of classes, you can also refer to DDIC types using LIKE (... comp11 LIKE ddic_type, ...). If you actually want to refer to an existing data object, but due to typing errors you inadvertently specify a name that exists as DDIC type, errors may be unavoidable.
• It is possible to create recursive structure references. These are components of a structured type that represent data references to the same structure in which they are defined (see the comp10 component in the struc_type example).

⬆️ back to top

Creating Structures

• To create a structure (i.e. a structured data object) in an ABAP program, you can use the DATA keyword.
• It works in the same way as the TYPES statement above.
• Unlike the TYPES statement, you can use the VALUE addition to set default values.

DATA: BEGIN OF struc,
        comp1 TYPE ...,
        comp2 TYPE ... VALUE ...,
        comp3 TYPE i VALUE 99,
        comp4 TYPE i VALUE IS INITIAL,  "Without the addition VALUE, or if IS INITIAL is specified, 
                                        "the content is initial.
        comp5 TYPE local_structured_type,
        ...,
      END OF struc.

Alternatively, you can use the following syntax. Similar to above, a chained statement may provide better readability.

DATA BEGIN OF struc.
  DATA comp1 TYPE ... .
  DATA comp2 TYPE ... VALUE ... .
... .
DATA END OF struc.

[!NOTE]

• The keywords CLASS-DATA and CONSTANTS can also be used to create structures. In principle, they represent special cases of the general statement shown above. See the ABAP Keyword Documentation for more information.
• Structures can also be created inline using DATA(...) or FINAL(...), as shown below.

⬆️ back to top

Creating Structures Using Existing Structured Types

"Local structured type 
TYPES: BEGIN OF struc_type,
         comp1 TYPE i,                            
         comp2 TYPE c LENGTH 5,   
       END OF struc_type.

"Creating a structure using a local structured type
DATA struc_1 TYPE struc_type.

"Creating structures based on globally available types from the DDIC
"Note: When referring to such types, you cannot provide start values for the individual components. 
DATA: struc_2 TYPE some_ddic_structure,
      struc_3 TYPE some_ddic_table,
      struc_4 TYPE some_cds_view.

"Structure based on a structured type that is available in the public
"visibility section of a global class
DATA struc_5 TYPE cl_some_class=>struc_type.

"Creating structures by referring to local data objects and internal table types
DATA: struc_6 LIKE struc_1,
      struc_7 LIKE LINE OF some_itab,
      struc_8 TYPE LINE OF some_itab_type.

⬆️ back to top

Creating Structures by Inline Declaration

• This is particularly useful for declaring data objects at the operand positions where you actually need them.
• In this way, you can avoid an extra declaration of the structure in different contexts.
• You can use the declaration operator using DATA(...). The FINAL declaration operator is used to create immutable variables.
• You can also create structures using the VALUE operator (and also fill them as shown below). Without specifying component values in the parentheses, you create an initial structure.

"Structures created inline instead of an extra declared variable
DATA struc_9 LIKE struc_1.
struc_9 = struc_1

"Type is derived from the right-hand structure; the content of struc is assigned, too.
DATA(struc_10) = struc_1.
FINAL(struc_11) = struc_9.

"Using the VALUE operator
"A structure declaration as follows (without providing component 
"value assignments) ...
DATA(struc_a) = VALUE struc_type( ).

"... is similar to the following declaration.
DATA struc_b TYPE struc_type.

"Structures declared inline instead of an extra declared variable

"Example: SELECT statement
"Extra declaration
DATA struc_12 TYPE zdemo_abap_fli.

SELECT SINGLE *
  FROM zdemo_abap_fli
  WHERE carrid = 'LH'
  INTO @struc_12.

"Inline declaration
SELECT SINGLE *
  FROM zdemo_abap_fli
  WHERE carrid = 'LH'
  INTO @DATA(struc_13).

"Example: Loop over an internal table
DATA itab TYPE TABLE OF zdemo_abap_fli WITH EMPTY KEY.
... "itab is filled

"Extra declaration
DATA wa_1 LIKE LINE OF itab.

LOOP AT itab INTO wa_1.
  ...
ENDLOOP.

"Inline declaration
LOOP AT itab INTO DATA(wa_2).
  ...
ENDLOOP.

⬆️ back to top

Creating Constant and Immutable Structures

• Constant structures can be created with the ... BEGIN OF ... END OF ... additions. Their values cannot be changed.
• As shown above, the FINAL declaration operator is used to create immutable variables.

CONSTANTS: BEGIN OF const_struct,
             num TYPE i VALUE 123,
             str TYPE string VALUE `ABAP`,
             n3  TYPE n LENGTH 3 VALUE '000',
             c5  TYPE c LENGTH 5 VALUE 'abcde',
           END OF const_struct.

DATA(num) = const_struct-num.
"const_struct-num = 456.

TYPES struct_type LIKE const_struct.
FINAL(final_struct) = VALUE struct_type( num = 987 str = `hello` n3 = '123' c5 = 'xyz' ).

DATA(num_from_final) = final_struct-num.
"final_struct-num = 1.

SELECT * FROM zdemo_abap_carr INTO TABLE @DATA(itab).

"The work area is specified as immutable variable. The variable's content cannot be changed
"in the loop, however, the variable is exchanged with the next loop pass.
LOOP AT itab INTO FINAL(wa).
  DATA(carrid) = wa-carrid.
  "wa-carrid = 'XY'.
ENDLOOP.

⬆️ back to top

Creating Enumerated Structures

Find more information on enumerated types in the Data Types and Data Objects cheat sheet.

"When creating enumerated types, an enumerated structure can optionally be declared in 
"the context of the type declaration.
"A component of an enumerated structure: An enumerated constant that exists as a component
"of a constant structure, not as a single data object.
TYPES basetype TYPE i.
TYPES: BEGIN OF ENUM t_enum_struc STRUCTURE en_struc BASE TYPE basetype,
         a VALUE IS INITIAL,
         b VALUE 1,
         c VALUE 2,
         d VALUE 3,
       END OF ENUM t_enum_struc STRUCTURE en_struc.

DATA(enum_comp) = en_struc-b.

DATA(conv_enum_comp) = CONV basetype( en_struc-b ).
ASSERT conv_enum_comp = 1.

⬆️ back to top

Creating Anonymous Structures

Using the instance operator NEW and CREATE DATA statements, you can create anonymous data objects, such as anonymous structures. The NEW addition of the INTO clause of an ABAP SQL SELECT statement also creates an anonymous data object. You can access the components or the entire data objects by dereferencing. For more information, refer to the Dynamic Programming and Constructor Expressions cheat sheets.

"Without assigning component values in the parentheses, the anonymous
"structure is initial.
DATA(struc_ref_a) = NEW struc_type( ).

DATA struc_ref_b TYPE REF TO DATA.
struc_ref_b = NEW struc_type( ).

"Multiple syntax options are available for CREATE DATA
"statements. See the cheat sheets mentioned.
CREATE DATA struc_ref_b TYPE struc_type.

DATA struc_ref_c TYPE REF TO struc_type.
"Implicit data type definition
CREATE DATA struc_ref_c.

"NEW addition of the INTO clause of an ABAP SQL SELECT statement
SELECT SINGLE carrid, carrname
 FROM zdemo_abap_carr
 WHERE carrid = char`LH`
 INTO NEW @DATA(struc_ref_d).

⬆️ back to top

Variants of Structures

Depending on the component type, the structure can be a flat structure, a nested structure, or a deep structure.

• Flat structures contain only elementary types that have a fixed length, that is, there are no internal tables, reference types or strings as components. Nesting does not matter in this context. Even a nested structure is considered flat unless a substructure contains a deep component.

  DATA: BEGIN OF struc,
          comp1 TYPE i,
          comp2 TYPE c LENGTH 15,
          comp3 TYPE p LENGTH 8 DECIMALS 2,
          ...,
        END OF struc.
  

• Nested structures: At least one component of a structure is a substructure, that is, it refers to another structure. The following example has multiple substructures.

  DATA: BEGIN OF address_n,
          BEGIN OF name,
            title   TYPE string VALUE `Mr.`,
            prename TYPE string VALUE `Duncan`,
            surname TYPE string VALUE `Pea`,
          END OF name,
          BEGIN OF street,
            name TYPE string VALUE `Vegetable Lane`,
            num  TYPE string VALUE `11`,
          END OF street,
          BEGIN OF city,
            zipcode TYPE string VALUE `349875`,
            name    TYPE string VALUE `Botanica`,
          END OF city,
      END OF address_n.
  

• Deep structures: Contain at least one internal table, reference type, or string as a component.

  DATA: BEGIN OF address_d,
          name    TYPE string VALUE `Mr. Duncan Pea`, 
          street  TYPE string VALUE `Vegetable Lane 11`, 
          city    TYPE string VALUE `349875 Botanica`, 
          details TYPE TABLE OF some_table WITH EMPTY KEY, 
        END OF address_d.
  

  Although the following structure looks quite simple, it is not a flat structure, but a deep structure, because it contains strings.

  DATA: BEGIN OF address,
          name   TYPE string VALUE `Mr. Duncan Pea`,
          street TYPE string VALUE `Vegetable Lane 11`,
          city   TYPE string VALUE `349875 Botanica`,
        END OF address.
  

[!NOTE]

• The data types of DDIC types are all flat (not nested) structures. Exception: Components of type string can be contained.
• Work areas of ABAP SQL statements cannot contain any deep components other than strings among others.
• Especially for assignments and comparisons of deep structures, the compatibility of the source and target structure must be taken into account.

⬆️ back to top

Accessing (Components of) Structures

• Structures can be accessed as a whole. You can also address the individual components of structures at the appropriate operand positions.
• To address the components, use the structure component selector -.
• For variables with reference to a structured data object, the object component selector -> can be used: ...dref->comp .... The following syntax also works, but is less convenient: ... dref->*-comp ....
• ADT and the ABAP Editor provide code completion for structure components after the component selectors.

"Addressing components via the structure component selector
... struc-comp1 ...
... struc-comp2 ...
... struc-comp3 ...

"Examples for addressing the whole structure and individual components
IF struc IS INITIAL. 
  ...
ENDIF.

IF struc-comp1 = 1. 
  ...
ENDIF.

DATA(complete_struc) = struc.
DATA(comp_value) = struc-comp2.

"Type and data declarations
TYPES: type_1 TYPE structured_type-comp1,
       type_2 LIKE struc-comp1.

DATA: var_1 TYPE structured_type-comp1,
      var_2 LIKE struc-comp1.

"Variables with reference to a structured data object
DATA ref_struc_1 TYPE REF TO structured_type.
ref_struc_1 = NEW #( ).
"Excursion: Creating a reference variable using inline declaration
DATA(ref_struc_2) = NEW structured_type( ).

... ref_struc_1->comp1 ...
... ref_struc_1->*-comp1 ...  "Using the dereferencing operator
... ref_struc_2->comp2 ... 
... ref_struc_2->*-comp2 ...  "Using the dereferencing operator

Nested components can be addressed using chaining:

... struc-substructure-comp1 ...
... address_n-name-title ...

[!NOTE]
There are syntax options for dynamically accessing structure components. See the Dynamic Porgramming cheat sheet.

⬆️ back to top

ASSIGN Statements

"A field symbol is set using an assignment of a memory area to the 
"field symbol by ASSIGN statements.
"Particularly, field symbols and ASSIGN statements are supporting elements for 
"dynamic programming. ASSIGN statements have multiple additions. Find more information 
"and examples in the Dynamic Programming cheat sheet.

TYPES: BEGIN OF s,
         comp1 TYPE i,
         comp2 TYPE c LENGTH 3,
         comp3 TYPE n LENGTH 5,
         comp4 TYPE string,
       END OF s.

DATA(demo_struc) = VALUE s( comp1 = 1 comp2 = 'abc' comp3 = '12345' comp4 = `ABAP` ).

"Defining a field symbol
FIELD-SYMBOLS <a> TYPE s.

"Assigning the entire structure to a field symbole
ASSIGN demo_struc TO <a>.

"Accessing a structure component via the field symbol
DATA(comp1) = <a>-comp1.

"Field symbol declared inline
"Note: The typing depends on the memory area specified. In this case,
"the field symbol <fd> has the structured type s. This is valid for static
"assignments. In case of dynamic assignments, the type is the generic type
"data.
ASSIGN demo_struc TO FIELD-SYMBOL(<b>).
comp1 = <b>-comp1.

"Accessing components of structures by assigning components to field symbols
"The field symbols is typed with the generic type data so that all components,
"which have random types, can be assigned.
FIELD-SYMBOLS <d> TYPE data.
ASSIGN demo_struc-comp1 TO <d>.
ASSIGN demo_struc-comp2 TO <d>.
ASSIGN demo_struc-comp3 TO <d>.
ASSIGN demo_struc-comp4 TO <d>.

"Accessing structures and components dynamically
"Note: In case of dynamic assignments, ...
"- sy-subrc is set.
"- the type of field symbols declared inline is the generic type data.
ASSIGN ('DEMO_STRUC') TO FIELD-SYMBOL(<c>).
ASSERT sy-subrc = 0.

"Using ASSIGN COMPONENT statements
"It is recommended to use the newer syntax, which specifies the component 
"selector followed by a data object in a pair of parentheses.
"After COMPONENT (and in the parentheses in the newer syntax), a character-like 
"or numeric data object is expected.
DATA(some_comp) = `COMP2`.
ASSIGN COMPONENT some_comp OF STRUCTURE demo_struc TO <d>.
ASSERT sy-subrc = 0.

some_comp = `COMP5`.
ASSIGN COMPONENT some_comp OF STRUCTURE demo_struc TO <d>.
ASSERT sy-subrc = 4.

"Numeric data objects
"The data object is implicitly converted to type i (if required) and
"interpreted as the position of the component in the structure.
ASSIGN COMPONENT 1 OF STRUCTURE demo_struc TO <d>.
ASSERT sy-subrc = 0.

ASSIGN COMPONENT 5 OF STRUCTURE demo_struc TO <d>.
ASSERT sy-subrc = 4.

"0 means that the entire structure is assigned
ASSIGN COMPONENT 0 OF STRUCTURE demo_struc TO <d>.
ASSERT sy-subrc = 0.
ASSERT <d> = demo_struc.

"Newer syntax using the component selector followed by content in
"a pair of parentheses.
ASSIGN demo_struc-('COMP4') TO <d>.
ASSIGN demo_struc-(some_comp) TO <d>.
ASSIGN demo_struc-(3) TO <d>.
ASSIGN demo_struc-(0) TO <d>.

"Iterating across all structure components
DO.
  ASSIGN demo_struc-(sy-index) TO <d>.
  IF sy-subrc <> 0.
    EXIT.
  ENDIF.
ENDDO.

⬆️ back to top

Populating Structures

You can copy the content of a structure to another using the assignment operator =. In the following example, it is assumed that the target and source structures are of compatible types. In general, note that special conversion and comparison rules apply to value assignments involving structures.

some_struc = another_struc.

"When creating a new structure by inline declaration, the type of
"the right-hand structure is derived and the content is assigned.

DATA(struc_inl) = some_struc.

To assign values to individual structure components, use the component selector.

TYPES: BEGIN OF addr_struc,
        name   TYPE string,
        street TYPE string,
        city   TYPE string,
       END OF addr_struc.

DATA address TYPE addr_struc.

address-name   = `Mr. Duncan Pea`.
address-street = `Vegetable Lane 11`.
address-city   = `349875 Botanica`.

⬆️ back to top

Using the VALUE Operator

• The VALUE operator can be used to construct the content of complex data objects such as structures or internal tables.
• It is particularly useful because assigning values by addressing the structure components individually can be very cumbersome, especially when assigning values to structure components at the operand position.
• If the type of the operand can be inferred implicitly, the # character can be used used before the parentheses. Otherwise, the type must be specified explicitly.
• The VALUE operator and inline declarations can be used to create and populate structures in one go.
• Note that there are special conversion and comparison rules for structures. See the ABAP Keyword Documentation for more details.

"# used: type of the operand can be implicitly derived
address = VALUE #( name   = `Mr. Duncan Pea`
                   street = `Vegetable Lane 11`
                   city   = `349875 Botanica` ).

"Declaring a structure inline
"Type used explicitly: type of the operand cannot be implicitly derived
DATA(addr) = VALUE addr_struc( name   = `Mr. Duncan Pea`
                               street = `Vegetable Lane 11`
                               city   = `349875 Botanica` ).


"Using the BASE addition to retain existing component values
addr = VALUE #( BASE addr street = `Some Street 1` ).
*NAME              STREET           CITY           
*Mr. Duncan Pea    Some Street 1    349875 Botanica

"Without the BASE addition, the components are initialized
addr = VALUE #( street = `Another Street 2` ).
*NAME       STREET              CITY   
*           Another Street 2           

"Nesting value operators
TYPES: BEGIN OF struc_nested,
        a TYPE i,
        BEGIN OF nested_1,
          b TYPE i,
          c TYPE i,
        END OF nested_1,
        BEGIN OF nested_2,
          d TYPE i,
          e TYPE i,
        END OF nested_2,
      END OF struc_nested.

DATA str_1 TYPE struc_nested.

str_1 = VALUE #( a        = 1 
                 nested_1 = VALUE #( b = 2 c = 3 ) 
                 nested_2 = VALUE #( d = 4 e = 5 ) ).

"Inline declaration
"Component a is not specified here, i.e. its value remains initial.
DATA(str_2) = VALUE struc_nested( nested_1 = VALUE #( b = 2 c = 3 ) 
                                  nested_2 = VALUE #( d = 4 e = 5 ) ).

"Apart from the VALUE operator, the NEW operator can be used to create
"a data reference variable (and populate the structure)
DATA(str_ref) = NEW struc_nested( a        = 1 
                                  nested_1 = VALUE #( b = 2 c = 3 ) 
                                  nested_2 = VALUE #( d = 4 e = 5 ) ).

⬆️ back to top

Using the NEW Operator

Using the instance operator NEW, you can create anonymous data objects, such as anonymous structures. You can access the components or the entire data objects by dereferencing. For more information, refer to the Dynamic Programming and Constructor Expressions cheat sheets.

"Creating a data reference variable 
DATA addr_ref1 TYPE REF TO addr_struc.

"Populating the anonymous structure
addr_ref1 = NEW #( name   = `Mr. Duncan Pea`
                   street = `Vegetable Lane 11`
                   city   = `349875 Botanica` ).

addr_ref1->name = `Mrs. Jane Doe`.     

"Declaring an anonymous structure/a data reference variable inline    
DATA(addr_ref2) = NEW addr_struc( name   = `Mr. Duncan Pea`
                                  street = `Vegetable Lane 11`
                                  city   = `349875 Botanica` ).

addr_ref2->* = VALUE #( BASE addr_ref2->* name = `Mr. John Doe` ).

⬆️ back to top

Using the CORRESPONDING Operator and MOVE-CORRESPONDING Statements

• You can use statements with MOVE-CORRESPONDING and the CORRESPONDING operator to assign values to structure components, especially when assigning values from a source structure to a target structure which have incompatible types and/or differently named components.
• Both are used to assign identically named components of structures to each other.
• The syntax also works for structures of the same type.
• Also note the special conversion and comparison rules for structures in this context.

[!NOTE]

• The CL_ABAP_CORRESPONDING system class is available for making assignments. See the ABAP Keyword Documentation for the details.
• The INTO clause of ABAP SQL statements has the CORRESPONDING addition. There, the following basic rule applies, which affects the value assignment: Without the CORRESPONDING ... addition, column names do not matter, only the position. With the CORRESPONDING ... addition, the position of the columns does not matter, only the name. See examples in the ABAP SQL cheat sheet.

The following examples demonstrate the value assignment using MOVE-CORRESPONDING statements and the CORRESPONDING operator with various additions. The focus is on flat structures only.

"Moves identically named components; content in other components
"of the targets structure are kept.
MOVE-CORRESPONDING struc TO diff_struc.

"Initializes target structure; moves identically named components
diff_struc = CORRESPONDING #( struc ).

"Same effect as the first MOVE-CORRESPONDING statement;
"addition BASE keeps existing content
diff_struc = CORRESPONDING #( BASE ( diff_struc ) struc ).

"MAPPING addition: Specifying components of a source structure that are
"assigned to the components of a target structure in mapping
"relationships.
diff_struc = CORRESPONDING #( BASE ( diff_struc ) struc MAPPING comp1 = compa ).

"EXCEPT addition: Excluding components from the assignment.
diff_struc = CORRESPONDING #( BASE ( diff_struc ) struc EXCEPT comp1 ).

Value assignments in deep structures

• In the context of deep structures, there are additional syntax variants available for MOVE-CORRESPONDING statements and the CORRESPONDING operator.
• The following examples focus on internal tables as structure components. Check out the syntax in action in the executable example.

"Nonidentical elementary component types are kept in target
"structure which is true for the below MOVE-CORRESPONDING statements;
"existing internal table content is replaced by content of
"the source table irrespective of identically named components
MOVE-CORRESPONDING deep_struc TO diff_deep_struc.

"Existing internal table content is replaced but the value
"assignment happens for identically named components only.
MOVE-CORRESPONDING deep_struc TO diff_deep_struc EXPANDING NESTED TABLES.

"Existing internal table content is kept; table content of the source
"structure are added but the value assignment happens like the first
"MOVE-CORRESPONDING statement without further syntax additions.
MOVE-CORRESPONDING deep_struc TO diff_deep_struc KEEPING TARGET LINES.

"Existing internal table content is kept; table content of the source
"structure are added; the value assignment happens like the statement
"MOVE-CORRESPONDING ... EXPANDING NESTED TABLES.
MOVE-CORRESPONDING deep_struc TO diff_deep_struc EXPANDING NESTED TABLES KEEPING TARGET LINES.

"Target structure is initialized; the value assignment for an internal
"table happens irrespective of identically named components.
diff_deep_struc = CORRESPONDING #( deep_struc ).

"Target structure is initialized; the value assignment for an internal
"table happens for identically named components only.
diff_deep_struc = CORRESPONDING #( DEEP deep_struc ).

"Nonidentical elementary component types are kept in target structure;
"internal table content is replaced; there, the value assignment
"happens like using the CORRESPONDING operator without addition.
diff_deep_struc = CORRESPONDING #( BASE ( diff_struc ) deep_struc ).

"Nonidentical elementary component types are kept in target structure;
"internal table content is replaced; there, the value assignment
"happens like using the CORRESPONDING operator with the addition DEEP.
diff_deep_struc = CORRESPONDING #( DEEP BASE ( diff_struc ) deep_struc ).

"Nonidentical elementary component types are kept in target structure;
"internal table content is kept, too, and table content of the
"source structure are added; there, the value assignment
"happens like using the CORRESPONDING operator without addition.
diff_deep_struc = CORRESPONDING #( APPENDING BASE ( diff_struc ) deep_struc ).

"Nonidentical elementary component types are kept in target structure;
"internal table content is kept, too, and table content of the
"source structure are added; there, the value assignment
"happens like using the CORRESPONDING operator with the addition DEEP.
diff_deep_struc = CORRESPONDING #( DEEP APPENDING BASE ( diff_struc ) deep_struc ).

⬆️ back to top

Clearing Structures

You can reset individual components to their initial values and clear the entire structure using the CLEAR keyword. Note that FREE statements also deletes the content, but they also release the initially allocated memory. space.

CLEAR struc-component.

CLEAR struc.

"This statement additionally releases memory space.
FREE struc.

"Note: An assignment using the VALUE operator without entries in the parentheses clears the structure. 
struc = VALUE #( ). 

"The same applies to data reference variables pointing to structures.
struc_ref = NEW #( ).

⬆️ back to top

Processing Structures

Structures are primarily used to process data from tables. In this context, structures often take on the role of a work area. The following code snippets cover only a selection. For more examples, see the cheat sheets about internal tables and ABAP SQL.

Structures in ABAP SQL Statements

The following code snippets cover a selection. Find more information and code snippets in the ABAP SQL cheat sheet.

"Creating a structure with a compatible type
DATA ls_fli1 TYPE zdemo_abap_fli.

SELECT SINGLE FROM zdemo_abap_fli
  FIELDS *
  WHERE carrid = 'LH'
  INTO @ls_fli1.

"Target structure declared inline
SELECT SINGLE FROM zdemo_abap_fli
  FIELDS *
  WHERE carrid = 'LH'
  INTO @DATA(ls_fli2).

SELECT SINGLE FROM zdemo_abap_fli
  FIELDS *
  WHERE carrid = 'AA'
  INTO CORRESPONDING FIELDS OF @ls_fli_diff.

SELECT SINGLE FROM @itab AS itab_alias
  FIELDS *
  WHERE ...
  INTO @DATA(ls_struc).
  "INTO CORRESPONDING FIELDS OF @some_existing_struc.

SELECT FROM zdemo_abap_fli
  FIELDS *
  WHERE carrid = 'AZ'
  INTO @DATA(ls_sel_loop).
      
  IF sy-subrc = 0.
    ...
  ENDIF.
ENDSELECT.

INSERT INTO dbtab VALUES @struc.

INSERT dbtab FROM @struc.

INSERT dbtab FROM @( VALUE #( comp1 = ... comp2 = ... ) ).

UPDATE dbtab FROM @struc.

UPDATE dbtab FROM @( VALUE #( comp1 = ... comp2 = ... ) ).

SELECT SINGLE *
  FROM dbtab
  WHERE ...
  INTO @DATA(wa).

UPDATE dbtab FROM @( VALUE #( BASE wa comp2 = ... comp4 = ... ) ).

MODIFY dbtab FROM @struc.

MODIFY dbtab FROM @( VALUE #( comp1 = ... comp2 = ... ) ).

DELETE dbtab FROM @struc.

DELETE dbtab FROM @( VALUE #( comp1 = ... ) ).

Structures in Statements for Processing Internal Tables

The following code snippets cover a selection. Find more information and code snippets in the Internal Tables cheat sheet.

READ TABLE itab INTO DATA(wa) INDEX 1.

READ TABLE itab ASSIGNING FIELD-SYMBOL(<fs>) INDEX 2.

READ TABLE itab REFERENCE INTO DATA(dref) INDEX 3.

DATA(ls_table_exp) = itab[ 3 ].

LOOP AT itab ASSIGNING FIELD-SYMBOL(<fs>).
  <fs>-comp1 = ...
  ...
ENDLOOP.

INSERT struc INTO TABLE itab.

APPEND struc TO itab.

MODIFY TABLE itab FROM struc.

⬆️ back to top

Including Structures

• INCLUDE TYPE and INCLUDE STRUCTURE statements are used in the context of local structures.
• Structured data objects and types created with ... BEGIN OF... END OF ... can use this syntax to include components of another structure, whether it is a locally defined or global structure, without creating substructures.
• INCLUDE TYPE can be used to include a structured type.
• You can use INCLUDE STRUCTURE to include a structure.

[!NOTE]

• They are not additions of ... BEGIN OF ... END OF ... but individual ABAP statements.
• If you use a chained statement with a colon to declare the structure, the inclusion of other structures with these statements interrupts the chained statement, that is, the components of the included structures are included as direct components of the superstructure.
• By using the optional AS addition and specifying a name, the included components can be addressed by this common name as if they were actually components of a substructure.
• The optional RENAMING WITH SUFFIX addition, followed by a name, gives the included components a suffix name to avoid naming conflicts with other components.

The following example shows how structured types and data objects are included in another structure. First, three structured types and a structured data object based on one of these types are created. Then, the types and the structure are included in the structured type address_type. As an excursion, Runtime Type Identification is used to retrieve the component names of created structured type address_type. Refer to the Getting Structured Type Information and Creating Structures at Runtime section. The executable example demonstrates a structure that includes other structures in this way.

TYPES: BEGIN OF name_type,
        title   TYPE string,
        prename TYPE string,
        surname TYPE string,
      END OF name_type,
      BEGIN OF street_type,
        name TYPE string,
        num  TYPE string,
      END OF street_type,
      BEGIN OF city_type,
        zipcode TYPE string,
        name    TYPE string,
      END OF city_type.

DATA city_struc TYPE city_type.

TYPES BEGIN OF address_type.
      INCLUDE TYPE name_type AS name.
      INCLUDE TYPE street_type AS street RENAMING WITH SUFFIX _street.
      INCLUDE STRUCTURE city_struc AS city RENAMING WITH SUFFIX _city.
TYPES END OF address_type.

DATA(component_names) = VALUE string_table( FOR wa IN CAST cl_abap_structdescr(
   cl_abap_typedescr=>describe_by_name( 'ADDRESS_TYPE' ) )->components ( CONV #( wa-name ) ) ).

*Content of COMPONENT_NAMES:
*TITLE         
*PRENAME       
*SURNAME       
*NAME_STREET   
*NUM_STREET    
*ZIPCODE_CITY  
*NAME_CITY     

⬆️ back to top

Generic Structured Types

Fully Generic Structured Types

• ANY STRUCTURE is a built-in generic ABAP type for fully generic structured types, similar to the ANY TABLE type for generic internal tables.
• It can be used to define formal parameters and field symbols.
• When defined, any structure can be assigned to field symbols and passed as actual parameters to formal parameters in procedure calls.
• Criteria for typing formal parameters with TYPE ANY STRUCTURE include:
  • Any optional parameter must have a default value specified.
  • You can specify importing and changing parameters with fully generic structured types if they are mandatory or defined as default by specifying a default structure.
  • The typing is not allowed for exporting and returning parameters, as well as for AMDP procedures.

The code snippet demonstrates that structures of any type can be assigned to a field symbol typed with ANY STRUCTURE.

FIELD-SYMBOLS <struc> TYPE ANY STRUCTURE.

TYPES: BEGIN OF ty_flat,
         comp1 TYPE c LENGTH 3,
         comp2 TYPE i,
         comp3 TYPE n LENGTH 5,
       END OF ty_flat,
       BEGIN OF ty_nested,
         comp1 TYPE c LENGTH 10,
         comp2 TYPE ty_flat,
       END OF ty_nested,
       BEGIN OF ty_deep,
         comp1 TYPE string,
         comp2 TYPE REF TO i,
         comp3 TYPE ty_nested,
         comp4 TYPE string_table,
       END OF ty_deep.

TYPES BEGIN OF ty_include.
INCLUDE TYPE ty_flat AS a.
INCLUDE TYPE ty_nested AS b RENAMING WITH SUFFIX _n.
TYPES END OF ty_include.

DATA: struc_flat    TYPE ty_flat,
      struc_nested  TYPE ty_nested,
      struc_deep    TYPE ty_deep,
      struc_include TYPE ty_include.

ASSIGN struc_flat TO <struc>.
ASSIGN struc_nested TO <struc>.
ASSIGN struc_deep TO <struc>.
ASSIGN struc_include TO <struc>.

⬆️ back to top

User-Defined Partially Generic Structured Types

• Using ANY STRUCTURE CONTAINING component_list, you can create partially generic structured types.
• "Partially" means these types impose constraints by predefining a set of components (component_list) following the CONTAINING addition. These components must be included as part of the structure, while other components may vary.
• component_list is not comma-separated and can consist of one or more components. The components can be either generic or non-generic.
• The syntax can be used to create local types with TYPES statements. Unlike the fully generic type ANY STRUCTURE, ANY STRUCTURE CONTAINING component_list cannot be used to declare field symbols and formal parameters directly. Instead, the locally declared types must be used.
• Rules:
  • When assigning non-generic structures to partially generic structures, the order of components in the partially generic structure is irrelevant.
  • The components of the non-generic structure must match the components of the generic structured type, meaning the component names must be identical, and their types must be compatible.
  • The non-generic structure can include more components than those specified in the partially generic type.
  • Two partially generic structured types are compatible if their constraints in the component list do not contradict in terms of component names and types.
  • Strict compatibility checks apply to components with elementary types, particularly concerning dictionary types. See here.

The following example demonstrates the local creation of partially generic structured types, which are used to define field symbols. The example highlights the following aspects:

• Various possible and impossible assignments of non-generic structures to field symbols typed with partially generic structured types
• Strict compatibility checks enforced for elementary dictionary types
• Include structure

"Non-generic structured types
TYPES: BEGIN OF ts_1,
         comp1 TYPE i,
         comp2 TYPE i,
       END OF ts_1,
       BEGIN OF ts_2,
         comp2 TYPE i,
         comp1 TYPE i,
       END OF ts_2,
       BEGIN OF ts_3,
         comp1 TYPE i,
         comp2 TYPE i,
         comp3 TYPE i,
         comp4 TYPE i,
         comp5 TYPE i,
       END OF ts_3,
       BEGIN OF ts_4,
         comp1 TYPE i,
         comp2 TYPE int8,
       END OF ts_4.

DATA: struc_1 TYPE ts_1,
      struc_2 TYPE ts_2,
      struc_3 TYPE ts_3,
      struc_4 TYPE ts_4.

"Locally declared partially generic structured types
TYPES: ts_gen_1 TYPE ANY STRUCTURE CONTAINING comp1 TYPE i
                                              comp2 TYPE i,
       ts_gen_2 TYPE ANY STRUCTURE CONTAINING comp2 TYPE i
                                              comp3 TYPE i.

FIELD-SYMBOLS <gs_1> TYPE ts_gen_1.
FIELD-SYMBOLS <gs_2> TYPE ts_gen_2.

*&---------------------------------------------------------------------*
*& Various assignments with partially generic structured types
*&---------------------------------------------------------------------*

"Identical component list with compatible types
ASSIGN struc_1 TO <gs_1>.

"Component list matches, but different order
ASSIGN struc_2 TO <gs_1>.

"Component list of partially generic structure matches,
"assigned structure contains additional components
ASSIGN struc_3 TO <gs_1>.

"--- ERROR ---
"Component names match, but there is a type incompatibility;
"no assignment possible
"ASSIGN struc_4 TO <gs_1>.

"--- ERROR ---
"Not matching component list; no assignment possible
"ASSIGN struc_1 TO <gs_2>.

"--- ERROR ---
"Not matching component list; no assignment possible
"ASSIGN struc_2 TO <gs_2>.

"Component list of partially generic structure matches,
"assigned structure contains additional components
ASSIGN struc_3 TO <gs_2>.

"--- ERROR ---
"Not matching component list; no assignment possible
"ASSIGN struc_4 TO <gs_2>.

*&---------------------------------------------------------------------*
*& Strict compatibility checks enforced for elementary dictionary types
*&---------------------------------------------------------------------*

"Strict checks regarding elementary dictionary types
"The example uses the dictionary types timn/tims.
TYPES: BEGIN OF ts_time,
         time TYPE t,
       END OF ts_time.
DATA struc_time TYPE ts_time.

"Creating local timn/tims types
"The SELECT statements are just used to have a self-contained
"example to have a type to refer to.
SELECT SINGLE
  FROM i_timezone
  FIELDS tims`123456` AS tims
  INTO @data(tims).

SELECT SINGLE
  FROM i_timezone
  FIELDS timn`123456` AS timn
  INTO @data(timn).

TYPES ty_tims like tims.
TYPES ty_timn like timn.

TYPES ts_gen_tims TYPE ANY STRUCTURE
  CONTAINING time type ty_tims.
FIELD-SYMBOLS <gs_tims> TYPE ts_gen_tims.

ASSIGN struc_time TO <gs_tims>.

TYPES ts_gen_timn TYPE ANY STRUCTURE
  CONTAINING time TYPE ty_timn.
FIELD-SYMBOLS <gs_timn> TYPE ts_gen_timn.

"--- ERROR ---
"No assignment possible due to type incompatibility
"ASSIGN struc_time TO <gs_timn>.

*&---------------------------------------------------------------------*
*& Example with include structure
*&---------------------------------------------------------------------*

TYPES: BEGIN OF ts_5,
         comp1 TYPE i,
       END OF ts_5,
       BEGIN OF ts_6.
       INCLUDE TYPE ts_5 AS substruct.
TYPES: comp2 TYPE i,
       comp3 TYPE i,
       END OF ts_6.

DATA struc_incl TYPE ts_6.

"comp1 can be accessed both by comp1 and substruct-comp1
struc_incl-comp1 = 1.
struc_incl-substruct-comp1 = 1.
struc_incl-comp2 = 2.
struc_incl-comp3 = 3.

TYPES ty_gts_5 TYPE ANY STRUCTURE CONTAINING substruct TYPE ts_5
                                            comp2 TYPE i.

FIELD-SYMBOLS <fts_5> TYPE ty_gts_5.
ASSIGN struc_incl TO <fts_5>.

"The structure is also compatible with the following generic
"structured type, as there is no special handling for component
"aliases, and it meets the constraints set by the generic structure,
"even though comp1 in the example is essentially specified twice.
TYPES ty_gts_6 TYPE ANY STRUCTURE CONTAINING substruct TYPE ts_5
                                             comp1 TYPE i.

FIELD-SYMBOLS <fts_6> TYPE ty_gts_6.
ASSIGN struc_incl TO <fts_6>.

Expand the following collapsible section for example code. To try it out, create a demo class named zcl_demo_abap, or reuse it if it already exists. Paste the code into it. If you choose a different class name, update the class name in the code snippet accordingly. After activation, choose F9 in ADT to execute the class. The example is set up to display output in the console. To explore the entire example, you have imported the ABAP cheat sheet GitHub repository since it uses some of its artifacts.

The following example contexts are included:

• Assigning non-generic structures to field symbols typed with partially generic structured types using ASSIGN statements, highlighting component definitions.
• Structure assignments that emphasize position-based and name-based movements.
• Structure assignments that emphasize enforced strict checks regarding elementary dictionary types.
• Assigning non-generic structures with included structures to partially generic structures.
• Using internal tables with a line type of partially generic structured type.
• Implementing an algorithm in three ways, emphasizing the benefits of using generic structured types: Processing a BDEF derived type with a static, dynamic, and generic implementation.
  • In each method, content from an internal table typed with a BDEF derived type is processed, returning only instances where the %control structure is not initial. Additionally, %cid_ref values are extracted and returned.
  • Different tables of type TABLE FOR UPDATE and TABLE FOR DELETE are created as demo input for the methods.
  • The method with static implementation is only valid for one specific type, with the importing parameter request_static typed as a specific BDEF derived type. This method can process only one demo table.
  • The method with the dynamic implementation includes the importing parameter request_dynamic. It is typed as any table, accepting all types of tables. The method also includes a generic WHERE clause to be passed. In the example, the demo WHERE clause checks for non-initial %control components (%control is not initial). The demo tables of type TABLE FOR UPDATE can be processed in the example since these types include %control. However, any other tables can be passed, too. An internal table typed with TABLE FOR DELETE is passed, but it raises an exception as it cannot be processed (it has no %control component). The same applies to a demo string table passed.
  • The generic method features an importing parameter request_generic, based on a partially generic structure type declared with ANY STRUCTURE CONTAINING. The constraint is: The structure must have %cid_ref and %control components. The example defines %control with the fully generic structure typed ANY STRUCTURE, allowing it to accept all kinds of %control structures of BDEF derived types. This typing guarantees type safety, ensuring that all BDEF derived types with %cid_ref and %control components are accepted. Other tables, such as the demo table typed with TABLE FOR DELETE and the string table, cannot be passed.

CLASS zcl_demo_abap DEFINITION
  PUBLIC
  FINAL
  CREATE PUBLIC .

  PUBLIC SECTION.
    INTERFACES if_oo_adt_classrun.
    METHODS constructor.
  PROTECTED SECTION.
  PRIVATE SECTION.
    TYPES: BEGIN OF ts_1,
             comp1 TYPE i,
             comp2 TYPE i,
           END OF ts_1,
           BEGIN OF ts_2,
             comp2 TYPE i,
             comp1 TYPE i,
           END OF ts_2,
           BEGIN OF ts_3,
             comp1 TYPE i,
             comp2 TYPE i,
             comp3 TYPE i,
             comp4 TYPE i,
             comp5 TYPE i,
           END OF ts_3,
           BEGIN OF ts_4,
             comp1 TYPE i,
             comp2 TYPE int8,
           END OF ts_4,
           gen_struc   TYPE ANY STRUCTURE CONTAINING %cid_ref TYPE abp_behv_cid
                                                     %control TYPE ANY STRUCTURE,
           request_tab TYPE INDEX TABLE OF gen_struc WITH FURTHER SECONDARY KEYS,
           cid_ref_tab TYPE SORTED TABLE OF abp_behv_cid WITH UNIQUE KEY table_line,
           der_update1 TYPE TABLE FOR UPDATE zdemo_abap_rap_draft_m,
           der_update2 TYPE TABLE FOR UPDATE zdemo_abap_rap_ro_m,
           der_update3 TYPE TABLE FOR UPDATE zdemo_abap_rap_ro_u,
           der_delete  TYPE TABLE FOR DELETE zdemo_abap_rap_ro_m.

    DATA: struc_1 TYPE ts_1,
          struc_2 TYPE ts_2,
          struc_3 TYPE ts_3,
          struc_4 TYPE ts_4,
          update1 TYPE der_update1,
          update2 TYPE der_update2,
          update3 TYPE der_update3,
          delete  TYPE der_delete.

    METHODS: assignments IMPORTING out TYPE REF TO if_oo_adt_classrun_out,
      move IMPORTING out TYPE REF TO if_oo_adt_classrun_out,
      assignments_ddic_types IMPORTING out TYPE REF TO if_oo_adt_classrun_out,
      include_structures IMPORTING out TYPE REF TO if_oo_adt_classrun_out,
      itabs_generic_struc_type IMPORTING out TYPE REF TO if_oo_adt_classrun_out,
      static_implementation IMPORTING out            TYPE REF TO if_oo_adt_classrun_out
                                      request_static TYPE der_update1
                            EXPORTING cid_ref_tab    TYPE cid_ref_tab
                            RETURNING VALUE(result)  TYPE REF TO data,
      dynamic_implementation IMPORTING out                     TYPE REF TO if_oo_adt_classrun_out
                                       request_dynamic         TYPE ANY TABLE
                                       VALUE(dyn_where_clause) TYPE string
                             EXPORTING cid_ref_tab             TYPE cid_ref_tab
                             RETURNING VALUE(result)           TYPE REF TO data,
      generic_implementation IMPORTING out             TYPE REF TO if_oo_adt_classrun_out
                                       request_generic TYPE request_tab
                             EXPORTING cid_ref_tab     TYPE cid_ref_tab
                             RETURNING VALUE(result)   TYPE REF TO data,
      populate_structures.
ENDCLASS.


CLASS zcl_demo_abap IMPLEMENTATION.
  METHOD if_oo_adt_classrun~main.

    out->write( |\n| ).
    out->write( `*********** assignments ***********` ).
    out->write( |\n| ).
    assignments( out ).
    out->write( |\n| ).
    out->write( `*********** assignments_ddic_types ***********` ).
    out->write( |\n| ).
    assignments_ddic_types( out ).
    out->write( |\n| ).
    out->write( `*********** move ***********` ).
    out->write( |\n| ).
    move( out ).
    out->write( |\n| ).
    out->write( `*********** include_structures ***********` ).
    out->write( |\n| ).
    include_structures( out ).
    out->write( |\n| ).
    out->write( `*********** itabs_generic_struc_type ***********` ).
    out->write( |\n| ).
    itabs_generic_struc_type( out ).
    out->write( |\n| ).
    out->write( `*********** static_implementation ***********` ).
    out->write( |\n| ).
    static_implementation( out = out
                           request_static = update1 ).
    "The following data objects cannot be passed with this approach.
    "static_implementation( out = out request_static = update2 ).
    "static_implementation( out = out request_static = update3 ).
    "static_implementation( out = out request_static = delete ).

    out->write( |\n| ).
    out->write( `*********** dynamic_implementation ***********` ).
    out->write( |\n| ).

    DATA(dyn_where_clause) = `%control is not initial`.

    dynamic_implementation( out = out
                            request_dynamic = update1
                            dyn_where_clause = dyn_where_clause ).

    out->write( |\n| ).

    dynamic_implementation( out = out
                            request_dynamic = update2
                            dyn_where_clause = dyn_where_clause ).

    out->write( |\n| ).

    dynamic_implementation( out = out
                            request_dynamic = update3
                            dyn_where_clause = dyn_where_clause ).

    out->write( |\n| ).

    "The following example passes a BDEF derived type. However, it does
    "not contain the %control component, so the WHERE clause will not be met.
    dynamic_implementation( out = out
                            request_dynamic = delete
                            dyn_where_clause = dyn_where_clause ).

    out->write( |\n| ).

    "The following nonsense example method call passes a string table that
    "cannot be processed.
    DATA(str_tab) = VALUE string_table( ( `abc` ) ( `def` ) ).

    dynamic_implementation( out = out
                            request_dynamic = str_tab
                            dyn_where_clause = dyn_where_clause ).

    out->write( |\n| ).
    out->write( `*********** generic_implementation ***********` ).
    out->write( |\n| ).
    generic_implementation( out = out
                            request_generic = update1 ).
    out->write( |\n| ).
    generic_implementation( out = out
                            request_generic = update2 ).
    out->write( |\n| ).
    generic_implementation( out = out
                            request_generic = update3 ).
    "The following data objects cannot be passed with this approach.
    "generic_implementation( out = out request_generic = str_tab ).
    "generic_implementation( out = out request_generic = delete ).

  ENDMETHOD.

  METHOD assignments.

    populate_structures( ).

    TYPES: ts_gen_1 TYPE ANY STRUCTURE CONTAINING comp1 TYPE i
                                                  comp2 TYPE i,
           ts_gen_2 TYPE ANY STRUCTURE CONTAINING comp2 TYPE i
                                                  comp3 TYPE i.

    FIELD-SYMBOLS <gs_1> TYPE ts_gen_1.
    FIELD-SYMBOLS <gs_2> TYPE ts_gen_2.

    "Identical component list with compatible types
    ASSIGN struc_1 TO <gs_1>.

    out->write( data = <gs_1> name = `<gs_1>` ).

    "Component list matches, but different order
    ASSIGN struc_2 TO <gs_1>.

    out->write( data = <gs_1> name = `<gs_1>` ).

    "Component list of partially generic structure matches,
    "assigned structure contains additional components
    ASSIGN struc_3 TO <gs_1>.

    out->write( data = <gs_1> name = `<gs_1>` ).

    "Component names match, but there is a type incompatibility;
    "no assignment possible
    "ASSIGN struc_4 TO <gs_1>.

    "Not matching component list; no assignment possible
    "ASSIGN struc_1 TO <gs_2>.

    "Not matching component list; no assignment possible
    "ASSIGN struc_2 TO <gs_2>.

    "Component list of partially generic structure matches,
    "assigned structure contains additional components
    ASSIGN struc_3 TO <gs_2>.

    out->write( data = <gs_2> name = `<gs_2>` ).

    "Not matching component list; no assignment possible
    "ASSIGN struc_4 TO <gs_2>.

  ENDMETHOD.

  METHOD move.

    populate_structures( ).

    out->write( data = struc_2 name = `struc_2` ).

    "Position-based movement
    struc_2 = struc_1.
    ASSERT struc_2-comp1 <> 3.
    ASSERT struc_2-comp2 <> 4.

    out->write( data = struc_2 name = `struc_2` ).

    FIELD-SYMBOLS <s1> TYPE data.
    FIELD-SYMBOLS <s2> TYPE data.
    ASSIGN struc_1 TO <s1>.
    ASSIGN struc_2 TO <s2>.

    out->write( data = <s1> name = `<s1>` ).
    out->write( data = <s2> name = `<s2>` ).

    <s2> = <s1>.

    out->write( data = <s1> name = `<s1>` ).
    out->write( data = <s2> name = `<s2>` ).

    TYPES ts_gen TYPE ANY STRUCTURE CONTAINING comp1 TYPE i.
    FIELD-SYMBOLS <s4> TYPE ts_gen.
    ASSIGN struc_1 TO <s4>.

    out->write( data = <s4> name = `<s4>` ).

    struc_2 = <s4>.

    out->write( data = struc_2 name = `struc_2` ).

    "Corresponding move
    populate_structures( ).

    ASSIGN struc_1 TO <s4>.

    MOVE-CORRESPONDING <s4> TO struc_2.

    out->write( data = struc_2 name = `struc_2` ).

    populate_structures( ).

    ASSIGN struc_1 TO <s4>.

    struc_2 = CORRESPONDING #( <s4> ).

    out->write( data = struc_2 name = `struc_2` ).
  ENDMETHOD.

  METHOD assignments_ddic_types.

    "Strict checks regarding elementary dictionary types
    "The example uses the dictionary types timn/tims.
    TYPES: BEGIN OF ts_time,
             time TYPE t,
           END OF ts_time.
    DATA(struc_time) = VALUE ts_time( time = CONV t( '123456' ) ).

    "Creating local timn/tims types
    "The nonsense SELECT statements with the typed literals are just
    "used to have a self-contained example for a type to refer to.
    SELECT SINGLE
      FROM i_timezone
      FIELDS tims`123456` AS tims
      INTO @DATA(tims).

    SELECT SINGLE
      FROM i_timezone
      FIELDS timn`123456` AS timn
      INTO @DATA(timn).

    TYPES ty_tims LIKE tims.
    TYPES ty_timn LIKE timn.

    TYPES ts_gen_tims TYPE ANY STRUCTURE
      CONTAINING time TYPE ty_tims.
    FIELD-SYMBOLS <gs_tims> TYPE ts_gen_tims.

    ASSIGN struc_time TO <gs_tims>.

    out->write( data = <gs_tims> name = `<gs_tims>` ).

    TYPES ts_gen_timn TYPE ANY STRUCTURE
      CONTAINING time TYPE ty_timn.
    FIELD-SYMBOLS <gs_timn> TYPE ts_gen_timn.

    "--- ERROR ---
    "No assignment possible due to type incompatibility
    "ASSIGN struc_time TO <gs_timn>.

  ENDMETHOD.

  METHOD populate_structures.
    struc_1 = VALUE #( comp1 = 1 comp2 = 2 ).
    struc_2 = VALUE #( comp2 = 3 comp1 = 4 ).
    struc_3 = VALUE #( comp1 = 5 comp2 = 6 comp3 = 7
                       comp4 = 8 comp5 = 9 ).
    struc_4 = VALUE #( comp1 = 10 comp2 = 11 ).
  ENDMETHOD.

  METHOD include_structures.

    TYPES: BEGIN OF s1,
             comp1 TYPE i,
           END OF s1,
           BEGIN OF s2.
             INCLUDE TYPE s1 AS substruc.
    TYPES: comp2 TYPE i,
             comp3 TYPE i,
           END OF s2.

    TYPES t_gs1 TYPE ANY STRUCTURE CONTAINING substruc TYPE s1
                                              comp2 TYPE i.

    DATA(struct_1) = VALUE s2( substruc-comp1 = 1
                               comp2 = 2
                               comp3 = 3 ).

    FIELD-SYMBOLS <fs1> TYPE t_gs1.

    ASSIGN struct_1 TO <fs1>.

    out->write( data = <fs1> name = `<fs1>` ).

    TYPES: BEGIN OF s3.
             INCLUDE TYPE s1 AS substruc RENAMING WITH SUFFIX _sub.
    TYPES:   comp2 TYPE i,
             comp3 TYPE i,
           END OF s3.

    DATA(struct_2) = VALUE s3( substruc-comp1 = 4
                               comp2 = 5
                               comp3 = 6 ).

    "Not compatible due to renaming with suffix
    "ASSIGN struct_3 TO <fs1>.

    TYPES t_gs2 TYPE ANY STRUCTURE CONTAINING comp2 TYPE i
                                              comp1_sub TYPE i.

    FIELD-SYMBOLS <fs2> TYPE t_gs2.

    ASSIGN struct_2 TO <fs2>.

    out->write( data = <fs2> name = `<fs2>` ).

    TYPES t_gs3 TYPE ANY STRUCTURE CONTAINING comp2 TYPE i
                                              substruc TYPE s1.

    FIELD-SYMBOLS <fs3> TYPE t_gs2.

    ASSIGN struct_2 TO <fs3>.

    out->write( data = <fs3> name = `<fs3>` ).
  ENDMETHOD.

  METHOD itabs_generic_struc_type.

    TYPES: BEGIN OF ts_5,
             comp_a TYPE i,
             comp_b TYPE c LENGTH 10,
             comp_c TYPE decfloat34,
             comp_d TYPE string,
           END OF ts_5,
           tab_type1 TYPE TABLE OF ts_5 WITH EMPTY KEY.

    TYPES: c10         TYPE c LENGTH 10,
           t_gs4       TYPE ANY STRUCTURE CONTAINING comp_b TYPE c10
                                               comp_c TYPE decfloat34
                                               comp_a TYPE i,
           tab_type_gs TYPE TABLE OF t_gs4 WITH EMPTY KEY.

    FIELD-SYMBOLS <tgs> TYPE tab_type_gs.

    DATA(itab1) = VALUE tab_type1(
    ( comp_a = 1 comp_b = 'a'
      comp_c = CONV decfloat34( '1.1' ) comp_d = `hello` )
    ( comp_a = 2 comp_b = 'b'
      comp_c = CONV decfloat34( '2.2' ) comp_d = `world` )
    ( comp_a = 3 comp_b = 'c'
      comp_c = CONV decfloat34( '3.3' ) comp_d = `ABAP` ) ).

    ASSIGN itab1 TO <tgs>.

    out->write( <tgs> ).

    TYPES: BEGIN OF ts_6,
             comp_a TYPE i,
             comp_b TYPE c LENGTH 10,
           END OF ts_6,
           tab_type2 TYPE TABLE OF ts_6 WITH EMPTY KEY.

    DATA(itab2) = VALUE tab_type2( ( comp_a = 4 comp_b = 'd' )
                                   ( comp_a = 5 comp_b = 'e' ) ).

    <tgs> = CORRESPONDING #( BASE ( <tgs> ) itab2 ).

    out->write( data = <tgs> name = `<tgs>` ).

    struc_1 = VALUE #( comp1 = 6 comp2 = 7 ).
    DATA itab3 LIKE TABLE OF struc_1 WITH EMPTY KEY.
    APPEND struc_1 TO itab3.

    <tgs> = CORRESPONDING #( BASE ( <tgs> )
      itab3 MAPPING comp_a = comp1 comp_b = comp2 ).

    out->write( data = <tgs> name = `<tgs>` ).

    LOOP AT <tgs> ASSIGNING FIELD-SYMBOL(<wa>).
      <wa>-comp_b = to_upper( <wa>-comp_b ).
      ASSIGN <wa>-('comp_d') TO FIELD-SYMBOL(<comp>).
      <comp> = to_upper( <comp> ).
    ENDLOOP.

    SORT <tgs> BY comp_a DESCENDING.

    out->write( data = <tgs> name = `<tgs>` ).

    READ TABLE <tgs> ASSIGNING FIELD-SYMBOL(<line>)
      WITH KEY comp_b = 'C'.

    IF <line> IS ASSIGNED.
      out->write( data = <line> name = `<line>` ).
    ENDIF.

    DATA(ref) = REF #( <tgs>[ 5 ] OPTIONAL ).

    out->write( data = ref->* name = `ref->*` ).
  ENDMETHOD.

  METHOD dynamic_implementation.

    CREATE DATA result LIKE request_dynamic.

    dyn_where_clause = cl_abap_dyn_prg=>escape_quotes( dyn_where_clause ).

    TRY.
        LOOP AT request_dynamic ASSIGNING FIELD-SYMBOL(<instance>)
         WHERE (dyn_where_clause).

          INSERT <instance> INTO TABLE result->*.
          INSERT <instance>-('%cid_ref') INTO TABLE cid_ref_tab.

        ENDLOOP.

        out->write( data = result->* name = `result->*` ).
        out->write( data = cid_ref_tab name = `cid_ref_tab` ).
      CATCH cx_sy_itab_dyn_loop INTO DATA(error).
        out->write( |Error: { error->get_text( ) }| ).
    ENDTRY.

  ENDMETHOD.

  METHOD generic_implementation.
    CREATE DATA result LIKE request_generic.

    LOOP AT request_generic ASSIGNING FIELD-SYMBOL(<instance>)
      WHERE %control IS NOT INITIAL.

      INSERT <instance> INTO TABLE result->*.
      INSERT <instance>-%cid_ref INTO TABLE cid_ref_tab.

    ENDLOOP.

    out->write( data = result->* name = `result->*` ).
    out->write( data = cid_ref_tab name = `cid_ref_tab` ).
  ENDMETHOD.

  METHOD static_implementation.
    CREATE DATA result LIKE request_static.

    LOOP AT request_static ASSIGNING FIELD-SYMBOL(<instance>)
      WHERE %control IS NOT INITIAL.

      INSERT <instance> INTO TABLE result->*.
      INSERT <instance>-%cid_ref INTO TABLE cid_ref_tab.

    ENDLOOP.

    out->write( data = result->* name = `result->*` ).
    out->write( data = cid_ref_tab name = `cid_ref_tab` ).
  ENDMETHOD.

  METHOD constructor.

    update1 = VALUE #(
      ( %cid_ref = `cid1` %control-num1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid2` %control-num1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid3` %control-num1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid4` %control-num1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid5` )
      ( %cid_ref = `cid6` ) ).

    update2 = VALUE #(
      ( %cid_ref = `cid7` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid8` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid9` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid10` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid11` )
      ( %cid_ref = `cid12` ) ).

    update3 = VALUE #(
      ( %cid_ref = `cid13` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid14` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid15` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid16` %control-field1 = if_abap_behv=>mk-on )
      ( %cid_ref = `cid17` )
      ( %cid_ref = `cid18` ) ).

    delete = VALUE #(
      ( %cid_ref = `cid19` )
      ( %cid_ref = `cid20` )
      ( %cid_ref = `cid21` )
      ( %cid_ref = `cid22` ) ).

  ENDMETHOD.
ENDCLASS.