如何在 Oracle 11g 中使用游标在表中插入数据？

Question

在这个过程中，我想插入 userid 值不同，就像在这个选择查询中我们插入所有数据一样。

T_ALLOCATIONCONFIG 表的内容：

 id   userid   customerid   subbatchtypeid
---- -------- ------------ ----------------
 1    29       10003        1
 1    30       10003        1

VW_BATCH 表的内容：

 batchid   batchname   subbatchtypeid   customerid   batchstatus_id
--------- ----------- ---------------- ------------ ----------------
 1         test        1                10003        1
 2         test1       1                10003        1
 3         test2       1                10003        1
 4         test3       1                10003        1
 5         test4       1                10003        1
 6         test5       1                10003        1
 7         test6       1                10003        1
 8         test7       1                10003        1
 9         test8       1                10003        1
 10        test9       1                10003        1
 11        test12      1                10003        1

我想要T_BATCHALLOCATION 表中的这种类型的结果：

 id   batchid   customerid  userid
---- --------- ----------- --------
 1    1         10003        29
 2    2         10003        29
 3    3         10003        29
 4    4         10003        29
 5    5         10003        29
 6    6         10003        30
 7    7         10003        30
 8    8         10003        30
 9    9         10003        30
 10   10        10003        30

假设我们有 10 条来自 vw_batch 的记录，所以我想在 batchallocation 表中插入 5 条带有 userid = 29 的记录和 5 条带有 userid = 30 的记录。

如何编写查询或存储过程来插入这种结果？

CREATE OR REPLACE PROCEDURE p_autoallocate_batches
AS
  BEGIN
    DECLARE
      --      v_coll_rules_id coll_rules.coll_rules_id%TYPE;
      v_customer_id t_customers.customer_id%TYPE;
      v_batchid     t_batch.batchid%TYPE;
      v_user_id     t_users.user_id%TYPE;
      c             types.cursortype;

      CURSOR c_aa IS
       SELECT b.customer_id,
              b.batchid,
              a.user_id
        FROM vw_batch b
        JOIN t_allocationconfig a
          ON a.customer_id = b.customer_id
         AND a.subbatchtypeid = b.subbatchtypeid
       WHERE b.batchstatus = 'New' AND a.isactive = 'Y';

    BEGIN
      -- Dbms_Output.Put_Line('Hello World');
      -- Insert Into Tmp (Column1) Values ('Started');

      OPEN c_aa;

      LOOP
        FETCH c_aa INTO v_customer_id, v_batchid, v_user_id;
        EXIT WHEN c_aa%NOTFOUND;

        IF (v_customer_id > 0 AND v_batchid > 0 AND v_user_id > 0) THEN
          INSERT INTO t_batchallocation (customer_id, batchid, user_id, created_by, created_date)
          VALUES (v_customer_id, v_batchid, v_user_id, 1, sysdate);

          UPDATE t_batch
          SET batchstatus_id = 2, modified_by = 1, modified_date = sysdate
          WHERE batchid = v_batchid;
        END IF;

        --Insert Into Tmp (Column1) Values ('Started123');

        COMMIT;
      END LOOP;

      CLOSE c_aa;

    END;
  END;

Answer 1

不需要使用游标来执行此操作 - 您可以像这样在两个 sql 语句中执行此操作（但您必须创建一个类型来保存插入的 batch_id）：

create type num_list as table of integer;

create or replace procedure p_autoallocate_batches
as
  v_batch_ids num_list;
begin
  insert into t_batchallocation (customer_id,
                                 batchid,
                                 user_id,
                                 created_by,
                                 created_date)
  select b.customer_id,
         b.batchid,
         a.user_id,
         1,
         sysdate
  from   vw_batch b
         inner join t_allocationconfig a on (a.customer_id = b.customer_id
                                             and a.subbatchtypeid = b.subbatchtypeid)
  where  b.batchstatus = 'New'
  and    a.isactive = 'Y'
  and    b.customer_id > 0
  and    b.batchid > 0
  and    a.user_id > 0
  returning batchid bulk collect into v_batch_ids;

  update t_batch
  set    batchstatus_id = 2, modified_by = 1, modified_date = sysdate
  where  batchid in (select * from table(v_batch_ids));

  commit;

end p_autoallocate_batches;
/

注意未经测试，因为您没有提供表定义或示例数据

另外，我不确定我是否回答了您最初的问题，因为我不明白。请更改您的问题，以在您的表格中提供一些示例数据以及您希望以表格格式插入的内容。

---

好的，根据您的问题中添加的附加信息，我认为以下内容将满足您的要求：

with t_allocationconfig as (select 1 id, 29 userid, 10003 customerid, 1 subbatchtypeid from dual union all
                            select 1 id, 30 userid, 10003 customerid, 1 subbatchtypeid from dual),
               vw_batch as (select 1 batchid, 'test' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 2 batchid, 'test1' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 3 batchid, 'test2' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 4 batchid, 'test3' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 5 batchid, 'test4' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 6 batchid, 'test5' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 7 batchid, 'test6' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 8 batchid, 'test7' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 9 batchid, 'test8' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 10 batchid, 'test9' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual union all
                            select 11 batchid, 'test12' batchname, 1 subbatchtypeid, 10003 customerid, 1 batchstatus_id from dual)
--end of mimicking your sample data - see SQL below:
select vb.batchid id, -- is id supposed to be derived using a sequence, perhaps?
       vb.batchid,
       vb.customerid,
       ta.userid
from   (select id,
               userid,
               customerid,
               subbatchtypeid,
               row_number() over (partition by id, customerid, subbatchtypeid order by userid) rn,
               count(*) over (partition by id, customerid, subbatchtypeid) cnt
        from   t_allocationconfig) ta
       inner join (select batchid,
                          batchname,
                          subbatchtypeid,
                          customerid,
                          batchstatus_id,
                          row_number() over (partition by customerid, subbatchtypeid order by batchid) rn,
                          count(*) over (partition by customerid, subbatchtypeid) cnt
                   from   vw_batch) vb on (ta.customerid = vb.customerid
                                           and ta.subbatchtypeid = vb.subbatchtypeid
                                           and ta.rn = ceil(vb.rn/ceil(vb.cnt/ta.cnt)));

        ID    BATCHID CUSTOMERID     USERID
---------- ---------- ---------- ----------
         1          1      10003         29
         2          2      10003         29
         3          3      10003         29
         4          4      10003         29
         5          5      10003         29
         6          6      10003         29
         7          7      10003         30
         8          8      10003         30
         9          9      10003         30
        10         10      10003         30
        11         11      10003         30

这意味着您的代码应该类似于以下内容：

-- make sure you have a global type:
create type num_list as table of integer;

create or replace procedure p_autoallocate_batches
as
  v_batch_ids num_list;
begin
  insert into t_batchallocation (customer_id,
                                 batchid,
                                 user_id,
                                 created_by,
                                 created_date)
  select vb.customerid,
         vb.batchid,
         ta.userid,
         1,
         sysdate
  from   (select id,
                 userid,
                 customerid,
                 subbatchtypeid,
                 row_number() over (partition by id, customerid, subbatchtypeid order by userid) rn,
                 count(*) over (partition by id, customerid, subbatchtypeid) cnt
          from   t_allocationconfig) ta
         inner join (select batchid,
                            batchname,
                            subbatchtypeid,
                            customerid,
                            batchstatus_id,
                            row_number() over (partition by customerid, subbatchtypeid order by batchid) rn,
                            count(*) over (partition by customerid, subbatchtypeid) cnt
                     from   vw_batch) vb on (ta.customerid = vb.customerid
                                             and ta.subbatchtypeid = vb.subbatchtypeid
                                             and ta.rn = ceil(vb.rn/ceil(vb.cnt/ta.cnt)))
  returning batchid bulk collect into v_batch_ids;

  update t_batch
  set    batchstatus_id = 2, modified_by = 1, modified_date = sysdate
  where  batchid in (select * from table(v_batch_ids));

  commit;

end p_autoallocate_batches;
/

注意您可能不得不使用 partition by 和 join 子句以确保数据被分组并连接到正确的列上 - 我猜测其背后的逻辑是什么，因为您没有说。希望您能够解构查询以查看它在做什么，并进行适当的修改！