ABAP-SQL基础知识

SQL语法

我们在编写ABAP4程序的时候,经常需要从TABLE中根据某些条件读取数据,读取数据最常用的方法就是通过SQL语法实现的。ABAP/4中可以利用SQL语法创建或读取TABLE，SQL语法分为DDL(DATA DEFINE LANGUAGE)语言和DML(DATA MULTIPULATION LANGUAGE)语言，DDL语言是指数据定义语言，例如CREATE等, DML语言是数据操作语言，例如SELECT, INSERT等语句。SQL语句有OPEN SQL语句和NATIVE SQL语句。 OPEN SQL语句不是标准SQL语句，是ABAP/4语言，利用OPEN SQL语句能在Databases 和 Command 之间产生一个BUFFER，所以它有一个语言转换的过程。而NATIVE SQL语句则是标准的SQL语句, 它直接针对Databases操作。

OPEN SQL

SELECT语句
语法格式:
SELECT <result> [INTO <target>] [FROM <source>] [WHERE <condition>]
               [GROUP BY <fields>] [ORDER BY <sort order>]
其中: <result>指定要抓取的栏位
      <target>将读取的记录存放在work area中
      <source>指定从那个TABLE中读取资料
      <condition>抓取资料的条件
      <fields>指定按那些栏位分组
      <sort order>排序的栏位及方式
相关的系统变量:
     SY-SUBRC = 0 表示读取数据成功
              <> 0 表示未找到符合条件的记录
     SY-DBLNT: 被处理过的记录的笔数.
相关的命令:
     EXIT. 退出循环.
     CHECK <logistic statement>.如果逻辑表达式成立，则继续执行，否则，开始下一次循环。
  
利用循环方式读取所有记录
SELECT ….ENDSELECT.是循环方式读取记录的。
     例如:
TABLES MARD. ''声明要使用的工作区
SELECT [DISTINCT] * FROM MARD WHERE MATNR = '3520421700'.
    <Statements>.
ENDSELECT.
(从MARD中抓取所有料号=3520421700的资料)
◆读取一笔资料
      Example:
               TABLES SPFLI.
               SELECT SINGLE * FROM SPFLI   
                     WHERE PLANT ='CHUNGLI' AND TEL='4526174'.
               WRITE: / SPFLI-COMPANY,SPFLI-PLANT,SPFLI-TEL.
◆将读取的记录放在work area中,并且加入Internal table 中.
格式有:
   ... INTO <work area>
   ... INTO CORRESPONDING FIELDS OF <work area> ''匹配赋值
   ... INTO (f1, ..., fn) 变量组.
   ... INTO TABLE <internal table>
   ... INTO CORRESPONDING FIELDS OF TABLE <internal table>
   ... APPENDING TABLE <internal table>
   ... APPENDING CORRESPONDING FIELDS OF TABLE <internal table>
举例一:
TABLES MARD.
DATA: BEGIN OF ITAB OCCURS 10,
              MATNR LIKE MARD-MATNR,
              WERKS LIKE MARD-WERKS,
              LGORT LIKE MARD-LGORT,
              LABST LIKE MARD-LABST,
        END OF ITAB.
SELECT MATNR WERKS LGORT LABST
              INTO CORRESPONDING FIELDS OF ITAB
              FROM MARD
              WHERE MATNR = '3520421700'.
       APPEND ITAB.
       CLEAR ITAB.
ENDSELECT.
(将读取的结果放在Internal table ITAB中)

DATA: BEGIN OF WA,
        LINE(240),
      END OF WA.
DATA NAME(10).
NAME = 'SPFLI'.
SELECT * FROM (NAME) INTO WA.
WRITE: / WA-LINE.
ENDSELECT.
数据库表名称 SPFLI 被赋给字符字段 NAME。SELECT 语句将所有的行从 SPFLI 中读到目标区 WA 中。在该示例中，WA 与 SPFLI 的结构并不相同，每一行都将自动地转换成字符字段.

举例二.
TABLES MARD.
SELECT MATNR MTART MAKTX INTO (t_matnr, t_mtart, maktx)
               FROM MARD
               WHERE MATNR = '3520421700'.
       <Statements>.
ENDSELECT.
(从MARD中抓取料号=3520421700的料号、类型和描述,放在变量t_matnr, t_mtart, maktx中)。

Example:
                 TABLES SPFLI.
                 DATA WA LIKE TABLES.
                 SELECT * FROM SPFLI INTO WA.
                    WRITE: / WA-COMPANY,WA-PLANT.
                 ENDSELECT.
                 逐笔写入WA 工作区中
举例三.
将读取的资料写入 Initial Table 中
语法:
        SELECT .. INTO TABLE <itab>  
        Example:
        TABLES SPFLI.
        DATA ITAB LIKE SPFLI OCCURS 10 WITH HEADER LINE.
        SELECT * FROM SPFLI INTO ITAB.
        一次读10笔(Initial Table的长度)记录存入 ITAB 中
       
        SELECT .. INTO TABLE <itab> PACKAGE SIZE <n>
        一次读取 <n> 笔记录至 <itab>中
        Example:
           TABLES SPFLI.
           DATA ITAB LIKE SPFLI OCCURS 10 WITH HEADER LINE.
           SELECT * FROM SPFLI INTO ITAB PACKAGE SIZE 5.
           一次读取 5 笔记录

exp :

TABLES SPFLI.
DATA: BEGIN OF WA,
         NUMBER TYPE I VALUE 1,
         CITYFROM LIKE SPFLI-CITYFROM,
         CITYTO   LIKE SPFLI-CITYTO,
      END OF WA.
SELECT * FROM SPFLI INTO CORRESPONDING FIELDS OF WA.
   WRITE: / WA-NUMBER, WA-CITYFROM, WA-CITYTO.
ENDSELECT.
输出如下所 示：

在该示例中，系统只将数据库表 SPFLI 中选定行的列 CITYFROM 和 CITYTO 传送到 WA 中。WA 中的组件 NUMBER 保持不变。
*-------------------------------------------------------------------------------------*

按指定的栏位排序
TABLES SBOOK.
SELECT * FROM SBOOK   WHERE CARRID = 'LH' AND
                            CONNID = '0400' AND
                            FLDATE = '19950228'
                            ORDER BY BOOKID ASCENDING.
   WRITE: / SBOOK-BOOKID,   SBOOK-CUSTOMID,
           SBOOK-CUSTTYPE, SBOOK-SMOKER,
           SBOOK-LUGGWEIGHT, SBOOK-WUNIT,
           SBOOK-INVOICE.
ENDSELECT.
(利用参数ORDER BY所指定的栏位排序)
*-------------------------------------------------------------------------------------*

◆ 抓取数据的条件关键字
(1) BETWEEN <g1> AND <g2>
例如: WHERE YEAR BETWEEN 1995 AND 2000.
(2) LIKE <g>
例如: WHERE NAME LIKE 'MIKE%'.
('%'是通配符号)
(3) IN (<g1>…<gn>)
是<g1>…<gn>里面的任意一个值即可.
例如: WHERE PLANT IN ('CHUNGLI', 'TAOYUAN','LIUTU').
(表示PLANT 只要是'CHUNGLI'或'TAOYUAN'或'LIUTU'都可以).
(4) ORDER BY 关键字
     指定排序的栏位或顺序
     (1). ..ORDER BY PRIMARY KEY.
        根据 PRIMARY KEY 递增排序
     (2)…ORDER BY <f1> [DESCENDING] <f2> [DESCENDING]
        Example:
                  SELECT * FROM IM ORDER BY PART .
*-------------------------------------------------------------------------------------*

INSERT 语句
◆从work area 加入到Internal Table中
格式: INSERT INTO <database> VALUES <work area>
例如:
DATA: BEGIN OF WA,
              CODE(6) TYPE C,
              NAME(30) TYPE C,
        END OF WA.
DATA: VEN LIKE WA OCCURS 10.
…
WA-CODE = '530120'.
WA-NAME = 'XINGDA ELECTRONICS CO.,LTD'.
INSERT INTO VEN VALUES WA .
如果work area的名称就是internal table的名称,可以直接写成:
      INSERT <internal table>
例如:
DATA: BEGIN OF WA OCCURS 10,
              CODE(6) TYPE C,
              NAME(30) TYPE C,
        END OF WA.
…
WA-CODE = '530120'.
WA-NAME = 'XINGDA ELECTRONICS CO., LTD'.
INSERT WA.
◆从另外一个Internal table中INSERT 资料
格式:
INSERT <itab1> FROM TABLE <itab2> [ACCEPTING DUPLICATE KEY]
将<itab2>中非NULL的资料加入<itab1>中,加上[ACCEPTING DUPLICATE KEY]能限制相同PRIMARY KEY不重复加入.

加入一笔记录至资料库

1.自 Work Area 工作区
   语法:
        INSERT INTO <database> VALUES <wa>
        Example:
                 TABLES SPFLI.
                 DATA WA   LIKE SPFLI.
                 WA-NO = '34051920'.
                 WA-COMPANY='DELTA'.
                 INSERT SPFLI VALUES WA.
                 将 ITAB 资料加入 SPFLI中, 也可写成 INSERT SPFLI FROM ITAB.
          
                  SPFLI-NO='34299876'.
                  SPFLI-COMPANY='HP'.
                  INSERT SPFLI FROM SPFLI.
                  将Work Area SPFLI中的资料加入资料库档案 SPFLI中
                  因Work Area SPFLI的结构与资料档 SPFLI一样, 所以也可
                  写成 INSERT SPFLI.

2.自 Internal Table
     语法:
           INSERT <database> FROM TABLE <itab> [ACCEPTING DUPLICATE KEY]
           将 <itab>中非 NULL的资料加入 <database>中, 加上 [ACCEPTING DUPLICATE
           KEY]能检查不加入有重覆primary key, 若有重覆则 SY-SUBRC 会传回 4
           Example:
                    TABLES SPFLI.
                    DATA ITAB LIKE SPFLI OCCURS 10 WITH HEADER LINE.
                    ITAB-NO = '34051920'.
                    ITAB-COMPANY = 'DELTA'.
                    APPEND ITAB.
                    ….
                    INSERT SPFLI FROM TABLE ITAB
                             ACCEPTING DUPLICATE KEY.
*-------------------------------------------------------------------------------------*

UPDATE 指令
   异动已存在的记录内容

1.使用 Primary Key
语法:
            UPDATE <database> FROM <wa>
            Example:
                     TABLES SPFLI.
                     DATA WA LIKE SPFLI.
                     WA-NO='34051920'.
                     WA-COMPANY='DELTA'.
                    UPDATE SPFLI FROM WA.
                    如 SPFLI 的 Primary Key是 NO, 则会找到 NO='34051920'的记录, 将其 COMPANY栏位异动为 DELTA

2.使用条件式
语法:
        UPDATE <database> SET < f1>=<values>… WHERE <condition>
        根据条件式异动符合条件式的记录
        Example:
                 UPDATE SPFLI SET NO ='34051920'
                                     COMPANY = 'DELTA'
                         WHERE TEL = '4526107'.
*-------------------------------------------------------------------------------------*

3.   MODIFY 语法
     MODIFY <internal table> [FROM <work area>].
根据 Primary Key 寻找资料档中符合的记录, 若找到则更新异动, 若找不到则新增记录
语法:
       MODIFY <database> FROM <wa>
       Example:
                WA-NO='34051920'.
                WA-COMPANY='DELTA'.
                MODIFY SPFLI FROM WA.
    
4.   DELETE 语法
     DELETE <internal table> [FROM <work area>].
或: DELETE <internal table> [WHERE <conditions>]
   删除资料档的记录

1.使用 Primary Key
语法:
            DELETE <database> FROM <wa>
            Example:
                     TABLES SPFLI.
                     DATA WA LIKE SPFLI.
                     WA-NO='34051920'.
                     WA-COMPANY='DELTA'.
                     DELETE SPFLI FROM WA.
                     如 SPFLI 的 Primary Key是 NO, 则会找到 NO='34051920'的记录, 找到后将此笔删除

2.使用条件式
语法:
        DELETE FROM <database> WHERE <condition>
        根据条件式删除符合条件式的记录
        Example:
                 DELETE FROM SPFLI WHERE AREA = 'AMERICAN'.
                                       
*-------------------------------------------------------------------------------------*

5. DATABASE CURSOR
   Database Cursor是一个资料库暂存区, 将经SELECT指令读取的记录存放至此暂存区, 再由此暂存区放至Work Area中, 可减少资料库读取的次数.
1.开启 Database Cursor
语法:
        OPEN CURSOR <c> FOR SELECT … WHERE <condition>
        Example:
                TABLES SPFLI.
                DATA: WA LIKE SPFLI,
                        C1 TYPE CURSOR.
                OPEN CURSOR C1 FOR SELECT * FROM SPFLI
                       WHERE AREA ='TAIWAN'.
2.读取 Database Cursor的资料存入 Work Area
语法:
      FETCH NEXT CURSOR <c> INTO <wa>
          Example:
                   FETCH NEXT CURSOR C1 INTO WA.
读取下一笔Cursor位置的资料存入WA, 如果已无资料可读, SY-SUBRC <>0.
关闭 Database Cursor
语法:
      CLOSE CURSOR <c>
      Example:
               CLOSE CURSOR C1.

*-------------------------------------------------------------------------------------*

COMMIT WORK & ROLLBACK WORK
要确定资料成功写入资料库，可使用COMMIT WORK指令，如:
   COMMIT WORK.
相反的，如果反悔要复原，可使用 ROLLBACK WORK, 可复原在上个COMMIT WORK指令之后的资料, 如:
   ROLLBACK WORK.
*-------------------------------------------------------------------------------------*
使用NATIVE SQL指令
语法格式:
EXEC SQL [PERFORMING <form>].
     <statements>
ENDEXEC.
举例一.
EXEC SQL.
CREATE TABLE AVERI_CLNT (
         CLIENT   CHAR(3) NOT NULL,
         ARG1     CHAR(3) NOT NULL,
         ARG2     CHAR(3) NOT NULL,
         FUNCTION CHAR(10) NOT NULL,
         PRIMARY KEY (CLIENT, ARG1, ARG2)
                          )
ENDEXEC.
举例二.
DATA: F1(3), F2(3), F3(3).
F3 = ' 1 '
EXEC SQL.
SELECT CLIENT, ARG1 INTO :F1, :F2 FROM AVERI_CLNT
         WHERE ARG2 = :F3
ENDEXEC

PERFORMING <form name>的使用:
如果NATIVE SQL的SELECT命令执行结果是抓到多笔记录，我们想要逐笔记录处理时，就用PERFORMING 参数，这个FORM能被逐次调用。如果想中止调用，就用EXIT FORM SQL结束调用。
例如:
DATA: F1(3), F2(3), F3(3).

F3 = '010'
EXEC SQL PERFORMING WRITE_AVERI_CLNT.
SELECT CLIENT, ARG1 INTO :F1, :F2 FROM AVERI_CLNT
         WHERE ARG2 = :F3
ENDEXEC.
FORM WRITE_AVERI_CLNT.
WRITE: / F1, F2.
ENDFORM.

DATA: BEGIN OF WA,
                        NAME(8),
                        AGE TYPE I,
                        END   OF WA.
                 DATA F1 TYPE I.
                 FI = 20.
                 EXEC SQL PERFORMING OUTPUT.
                    SELECT NAME,AGE INTO :WA FROM NAME_TABLE
                             WHERE AGE >= :F1.
                 ENDEXEC..
                 FORM OUTPUT.
                   WRITE: / WA-NAME,WA-AGE.
                 ENDFORM.

注意:
a. NATIVE SQL把TABLE中的MANDT(client)栏位当作一般栏位使用,所以在抓取资料时必须指定特定的Client;
b. NATIVE SQL中的SELECT语句没有CHECK权限的功能;
c. 在登入SAP R/3系统时,我们已经自动与Database连接,所以在执行NATIVE SQL时并不需要CONNECT语句;
d. 一条NATIVE SQL语句可以以分号;结束,一般情况下是以句号.结束.
e. 某些数据库系统对TABLE名字和FIELD名字有大小写区别,要正确书写.
f. 在NATIVE SQL中,双引号"不表示注释.