DYNAMIC_DATA

Create ABAP nested data types dynamically within your program

在程序内动态创建 ABAP 嵌套数据类型

USAGE 使用方法

Use abapgit to pull up the project code.

请使用 abapgit 拉取项目代码。

A runnable demo / smoke test report ZDYNAMIC_DATA_DEMO covering the examples below and the edge cases is included in the package. It is not required by the library and can be deleted after pulling.

包内附带一个可运行的演示/冒烟测试程序 ZDYNAMIC_DATA_DEMO，覆盖下文示例及边界场景；该程序不是类库运行所必需的，拉取后可以删除。

Usage 1 Generate through configuration fields 通过配置字段生成

  DATA: lr_type TYPE REF TO cl_abap_datadescr.
  DATA: lt_field_tab TYPE TABLE OF zdos_datadescr.
  DATA dyn_data TYPE REF TO data.
  FIELD-SYMBOLS: <fs_wa> TYPE any.
  
  INSERT VALUE #( fldname = 'hello'        fldtype = 'F' )        INTO lt_field_tab INDEX 1.
  INSERT VALUE #( fldname = 'author'       fldtype = 'F' )        INTO lt_field_tab INDEX 2.
  INSERT VALUE #( fldname = 'skills'       fldtype = 'T' )        INTO lt_field_tab INDEX 3.
  INSERT VALUE #( fldname = 'skills-name'  fldtype = 'F' )        INTO lt_field_tab INDEX 4.  "Using ‘-’ to connect superior and subordinate fields
  INSERT VALUE #( fldname = 'skills-level' fldtype = 'F' ) INTO lt_field_tab INDEX 5.
  
  lr_type = zcl_dynamic_object=>create_by_field_tab( field_tab = lt_field_tab type = 'S' ).
  
  TRY.
      CREATE DATA dyn_data TYPE HANDLE lr_type.
      IF dyn_data IS NOT INITIAL.
        ASSIGN dyn_data->*  TO <fs_wa>.
      ENDIF.
    CATCH cx_root INTO DATA(lr_exc).
      DATA(lv_message) = lr_exc->get_text( ).
      WRITE:/ lv_message.
  ENDTRY.

Usage 2 Generate via json 通过 json 生成

The JSON is parsed with the kernel sXML library, so type generation itself has no dependencies on other repositories (in the example below /ui2/cl_json is only used afterwards to fill the generated type with data).

JSON 解析基于内核的 sXML 库实现，类型生成本身不依赖任何其他仓库（下例中的 /ui2/cl_json 只是随后往生成的类型里填充数据时由调用方自己使用的）。

  DATA: lr_type TYPE REF TO cl_abap_datadescr.
  DATA dyn_data TYPE REF TO data.
  DATA json_data TYPE string.

  FIELD-SYMBOLS: <fs_wa> TYPE any.

  json_data = '{ "hello": "Hi! I am a nested data type created dynamically at runtime.",' &&
               '"author": { "name": "Jack Liang", "favLang": "ABAP", "github": "Jack-Liang/DYNAMIC_DATA" },' &&
               '"skills": [ { "name": "ABAP", "level": 95 }, { "name": "JSON", "level": 88 }, { "name": "Coffee", "level": 100 } ] }'.


  lr_type = zcl_dynamic_object=>create_by_json( json_data = json_data ).
  
  TRY.
      CREATE DATA dyn_data TYPE HANDLE lr_type.
      IF dyn_data IS NOT INITIAL.
        ASSIGN dyn_data->*  TO <fs_wa>.
        /ui2/cl_json=>deserialize( EXPORTING json = json_data  CHANGING data = <fs_wa> ).
      ENDIF.
    CATCH cx_root INTO DATA(lr_exc).
      DATA(lv_message) = lr_exc->get_text( ).
      WRITE:/ lv_message.
  ENDTRY.

You can also get a filled data object in one step — the JSON is parsed once and no JSON binder is needed on the caller side (booleans become X/initial, null stays initial). 也可以一步拿到填好值的数据对象——JSON 只解析一次，调用方不再需要任何 JSON 反序列化组件（布尔值填 X/初始，null 保持初始）：

  DATA dyn_data TYPE REF TO data.
  FIELD-SYMBOLS: <fs_wa> TYPE any.

  dyn_data = zcl_dynamic_object=>create_data_by_json( json_data   = json_data
                                                       infer_types = abap_true ).

  ASSIGN dyn_data->* TO <fs_wa>.  "<fs_wa> already contains the values 已填好值

Note: to handle the classic exceptions (INVALID_JSON etc.) use the CALL METHOD ... EXCEPTIONS form. 注：如需处理经典异常（INVALID_JSON 等），使用 CALL METHOD ... EXCEPTIONS 调用形式。

Usage 3 Created by basic type 通过基本类型创建

In Usage 1, if you pass a STRUF, such as SFLIGHT-CARRID or S_CARR_ID, to field_tab, the corresponding type is generated instead of the default String type. You can also specify the underlying ABAP type, such as ( INTTY = 'C' LENGT= 50 ) or ( INTTY = 'P' LENGT = 8 DECIM = 2 ), which will create a variable of the specified type and length.

在 Usage 1 中，如果给 field_tab 传入 STRUF，如 SFLIGHT-CARRID 或 S_CARR_ID，则会生成对应类型，而不是默认的 String 类型； 也可以指定基础的 ABAP 类型，例如 ( INTTY = 'C' LENGT = 50 ) 或者 ( INTTY = 'P' LENGT = 8 DECIM = 2 ), 这将会创建指定类型和长度的变量。

Usage 4 Name mapping 字段名映射

JSON keys longer than 30 characters (or otherwise invalid as ABAP component names) can be mapped to valid names via NAME_MAP. The mapping entries are matched case insensitively; if a mapped target itself violates the ABAP name rules (>30 characters, invalid characters, leading digit, -), INVALID_FIELD_NAME is raised. CREATE_DATA fills the values using the same mapping.

JSON key 超过 30 位（或不能作为 ABAP 组件名）时，可以通过 NAME_MAP 映射为合法名字；映射条目按大小写不敏感匹配。如果映射目标本身不合法（超长、非法字符、数字开头、含 -），同样抛 INVALID_FIELD_NAME。CREATE_DATA 填值使用同一映射。

  DATA(lt_map) = VALUE zcl_dynamic_object=>ty_name_map(
    ( json = 'aVeryLongJsonKeyNameThatExceedsThirtyCharactersXyz' abap = 'SHORT_NAME' ) ).

  lr_type = zcl_dynamic_object=>create_by_json( json_data   = json_data
                                                infer_types = abap_true
                                                name_map    = lt_map ).
  " components: SHORT_NAME instead of the 50 character key

Usage 5 Serialize back to JSON 反向序列化为 JSON

TO_JSON is the inverse of CREATE_DATA_BY_JSON: it serializes any generated (or arbitrary) dynamic data object back to JSON. Booleans (c length 1) become true/false, character-like initial values become null, numeric zeros stay numbers, dates/times are emitted in their internal format. NAME_MAP is applied in reverse (ABAP name -> JSON key); reference components raise UNSUPPORTED_TYPE; an unbound reference serializes as null.

TO_JSON 与 CREATE_DATA_BY_JSON 对称：把生成的（或任意动态的）数据对象序列化回 JSON。布尔（c 长度 1）输出 true/false，字符类初始值输出 null，数字零保持数字，日期/时间按内部格式输出。NAME_MAP 反向生效（ABAP 名 -> JSON 键）；引用组件抛 UNSUPPORTED_TYPE；未绑定的引用序列化为 null。

  lr_data = zcl_dynamic_object=>create_data_by_json( json_data   = json_data
                                                     infer_types = abap_true ).

  DATA(lv_json) = zcl_dynamic_object=>to_json( data     = lr_data
                                               name_map = lt_map ).
  " round trip: json -> data -> json

⚠️ Notion 重要说明

This is a new project. The author cannot guarantee that it will always run correctly. Therefore, please test it thoroughly before using it in a production environment.

If you pass INFER_TYPES = abap_true to CREATE_BY_JSON / CREATE_DATA_BY_JSON, the program will infer the possible data types based on json:

JSON	Inferred ABAP type
95	i (small integers, up to 9 digits become p otherwise)
88.5	p with length/decimals fitted to the literal
true / false	c length 1
"text" / null / 1e5	string
[] , ["a", "b"]	table of string
[10, 20]	table of i (consistent scalar items)

Errors are reported through the classic exceptions INVALID_JSON (unparsable JSON) and INVALID_FIELD_NAME (a key longer than 30 characters, containing characters other than A-Z 0-9 _, starting with a digit, or containing -, which is reserved as the hierarchy separator).

这是一个新项目。作者不能保证它总是正确运行。因此，在将其用于生产环境之前，请对其进行彻底的测试。

通过 json 生成时，给 CREATE_BY_JSON / CREATE_DATA_BY_JSON 传入 INFER_TYPES = abap_true，程序将根据 json 推断可能的数据类型，规则见上表；推断不出类型的值一律按 string 处理。解析失败会抛出经典异常 INVALID_JSON（无效 JSON）和 INVALID_FIELD_NAME（字段名超过 30 位、含 A-Z 0-9 _ 之外的字符、以数字开头，或包含层级分隔符 -）。

🌟 Looking forward to your suggestions 欢迎“一键三连”，欢迎增加新特性

Related Articles 相关文章：

1. 在程序内动态创建ABAP嵌套数据类型
2. abapGit 使用经验总结

This repository is included in dotabap

Thanks 鸣谢

This project also refers to some code from the following projects, and we hereby express our gratitude

JSON2ABAPType

ajson (the sXML based JSON parsing approach, MIT)