Oracle：如何 DEREF VARRAY 的 REF 和嵌套的 REF 表？

Question

我正在处理 Oracle 中的嵌套表和 VArray（澄清一下，这是为了在大学学习）。我有这个TABLE结构，我把数据放进去就一切OK了。

CREATE OR REPLACE TYPE Addendum AS OBJECT (
    idCode          VARCHAR2(10),
    signature_date  TIMESTAMP,
    --A VARRAY(5) OF REF STUDENT (STUDENT IF THE TYPE, TSTUDENT THE TABLE)
    studentlist     student_list,
    --A TABLE OF REF STAFF (STAFF IS THE TYPE, TSTAFF THE TABLE)
    staff_members   tStaff,
    pdi_manager     REF PDI
) FINAL;

CREATE TABLE Addendums OF Addendum
    NESTED TABLE staff_members STORE AS staffmembers_nt;
CREATE UNIQUE INDEX ADDENDUMS_IDCODE_UINDEX on ADDENDUMS (IDCODE);

CREATE OR REPLACE TYPE tAddendum AS TABLE OF Addendum;

在第一次尝试中，我执行了一个类似SELECT * FROM ADDENDUMS 的查询，它显示了这个东西：

0000000000,2019-09-01 00:00:00.000000,"{<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF}","{<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF}"
0000000001,2019-09-01 00:00:00.000000,"{<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF}","{<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF,<failed to load> oracle.sql.REF}"

我还尝试使用 DEREF 指令进行选择并明确引用列，其中：select ad.IDCODE, ad.SIGNATURE_DATE, DEREF(ad.STUDENTLIST), DEREF(ad.STAFF_MEMBERS), DEREF(ad.PDI_RESPONSABLE) from UOC.ADDENDUMS ad;，它仅显示正确取消引用的 PDI_MANAGER。

我可以用 VARRAY 和 NESTED TABLE 做什么？例如，SQL 的正确语法是什么（不是 this question 中的 PL/SQL）？

Answer 1

您可以使用表集合表达式CROSS JOINVARRAY（或嵌套表）然后DEREF(VALUE(collection_alias)).column_name 来获取值。

例如：

SELECT a.idcode,
       a.signature_date,
       DEREF(VALUE(s)).id AS id,
       DEREF(VALUE(s)).name AS name
FROM   ADDENDUMS a
       CROSS JOIN TABLE( a.studentlist ) s;

对于您的类型：

CREATE TYPE Student AS OBJECT(
  id NUMBER,
  name VARCHAR2(50)
);

CREATE TABLE students OF Student (
  CONSTRAINT students__id__pk PRIMARY KEY ( id )
);

CREATE TYPE student_list AS VARRAY(5) OF REF Student;

CREATE OR REPLACE TYPE Addendum AS OBJECT (
    idCode          VARCHAR2(10),
    signature_date  TIMESTAMP,
    --A VARRAY(5) OF REF STUDENT (STUDENT IF THE TYPE, TSTUDENT THE TABLE)
    studentlist     student_list
) FINAL;

CREATE TABLE Addendums OF Addendum;

还有一些示例数据：

BEGIN
  INSERT INTO Students ( id, name ) VALUES ( 1, 'AAA' );
  INSERT INTO Students ( id, name ) VALUES ( 2, 'BBB' );
  INSERT INTO Students ( id, name ) VALUES ( 3, 'CCC' );
  INSERT INTO Students ( id, name ) VALUES ( 4, 'DDD' );
  INSERT INTO Students ( id, name ) VALUES ( 5, 'EEE' );
  INSERT INTO Students ( id, name ) VALUES ( 6, 'FFF' );

  INSERT INTO Addendums ( idCode, signature_date, studentlist )
  VALUES (
    1,
    SYSTIMESTAMP,
    student_list(
      ( SELECT REF(s) FROM students s WHERE id = 2 ),
      ( SELECT REF(s) FROM students s WHERE id = 4 ),
      ( SELECT REF(s) FROM students s WHERE id = 1 )
    )
  );
END;
/

输出：

识别码 | SIGNATURE_DATE |身份证 |姓名 :----- | :---------------------------- | -: | :--- 1 | 01-OCT-20 05.01.00.699863 下午 | 2 | BBB 1 | 01-OCT-20 05.01.00.699863 下午 | 4 | DDD 1 | 01-OCT-20 05.01.00.699863 下午 | 1 | AAA

db小提琴here

（注意：没有为 TStaff 类型提供示例代码，所以我已经删掉了该部分；但是方法与 VARRAY 相同。）

---

在以后的 Oracle 版本（支持OFFSET m ROWS FETCH NEXT n ROWS 语法）中，您可以使用嵌套查询从VARRAY 中获取值；但是，这在 Oracle 11g 中不受支持：

SELECT a.idcode,
       a.signature_date,
       ( SELECT DEREF(VALUE(t)).name
         FROM   TABLE( a.studentlist ) t
         OFFSET 0 ROWS FETCH NEXT 1 ROW ONLY ) AS name1,
       ( SELECT DEREF(VALUE(t)).name
         FROM   TABLE( a.studentlist ) t
         OFFSET 1 ROWS FETCH NEXT 1 ROW ONLY ) AS name2,
       ( SELECT DEREF(VALUE(t)).name
         FROM   TABLE( a.studentlist ) t
         OFFSET 2 ROWS FETCH NEXT 1 ROW ONLY ) AS name3,
       ( SELECT DEREF(VALUE(t)).name
         FROM   TABLE( a.studentlist ) t
         OFFSET 3 ROWS FETCH NEXT 1 ROW ONLY ) AS name4,
       ( SELECT DEREF(VALUE(t)).name
         FROM   TABLE( a.studentlist ) t
         OFFSET 4 ROWS FETCH NEXT 1 ROW ONLY ) AS name5
FROM   ADDENDUMS a;

输出：

识别码 | SIGNATURE_DATE |姓名1 |名称2 |姓名3 |姓名4 |名称5 :----- | :------------------------ | :---- | :---- | :---- | :---- | :---- 1 | 01-OCT-20 17.09.08.790538 |血脑屏障 | DDD | AAA | 空 | 空

db小提琴here