Question

Create table t1_Fact ( Cur_date Date, Name varchar2(10), Event varchar2(50), Price Number(10,0), TAX Number(10,0), Flag Number );
Create table App_Fact ( Application_ID Number, Application_Name varchar2(100), Application_Price Number, Appliation_Tax Number, Flag Number );
Create table t2 ( Table_Name Varchar2(100), Table_Columns Varchar(100), Table_Measure varchar2(100), t3_columns varchar2(100), t3_measures varchar2(100), t3_Where_Clause varchar2(100) );
Create table t3 ( Cur_date Date, Name varchar2(10), Event varchar2(50), Application_ID Number, Application_Name varchar2(100), Application_Price Number, Appliation_Tax Number, Price Number(10,0), TAX Number(10,0), Flag Number );

表t2包含所有表名，每个源表和目标表的列名以及where子句条件。

在这里，我需要通过将事实表名称作为参数传递给分组，使用事实表，度量和where子句的列名将t3的数据插入到特定事实表中。

就像我们在过程中传递t1_Fact表一样，我们必须从t2中获取所有详细信息，并从t3中获取详细信息，然后插入t1_Fact中并将其保存到CSV文件中

我尝试了以下步骤，但是无法将数据插入到csv文件中。

注意：在以下过程中，我需要将表（表名）数据提取到q2中，我已将变量声明为recordsFetched VARCHAR2(3000 BYTE);。需要知道我们如何将记录声明为%rowtype的recordsFetched，而从光标获取表名时却不知道表名。我还能在UTL_FILE.PUT_Line()

中知道如何使用立即执行

Create or Replace Procedure CommonProcedure(sourceTableName IN VARCHAR2) Is
  tablename t2.Table_Name%TYPE;
  destcolumns t2.Table_Columns%TYPE;
  destMeasures t2.Table_Measure%TYPE;
  whereClause t2.t3_Where_Clause%TYPE;
  sourceColumns t2.t3_columns%TYPE;
  sourceMeasures t2.t3_measures%TYPE;
  q1 VARCHAR2(3000 BYTE);
q2 VARCHAR2(3000 BYTE);
 recordsFetched VARCHAR2(3000 BYTE);
  pathInfo VARCHAR2(3000 BYTE);

   v_file  UTL_FILE.FILE_TYPE;

 Cursor TableName Is  SELECT Table_Name FROM t2;

Begin


 Open c1;
 Loop
   Fetch TableName Into tablename;
   Exit When TableName%notfound;

    SELECT Table_Columns, Table_Measure, t3_columns, t3_measures INTO destcolumns,destMeasures,sourceColumns,sourceMeasures FROM t2 where Table_Name = tablename;

    q1 := 'INSERT INTO '||tablename||'('||destColumns||','||destMeasures||')'||
        ' ( SELECT '||sourceColumns||','||sourceMeasures||','||sourceTableName
          ||' FROM '||sourceTableName||' GROUP BY '||sourceColumns||')';

    Execute Immediate q1;

q2 := 'SELECT ' || destColumns || ',' || destMeasures ||' FROM ' || tablename || '';
           dbms_output.put_line(q2);
           EXECUTE IMMEDIATE  q2 into recordsfetched; --Getting an error here


         v_file := UTL_FILE.FOPEN('USER_DIR', destinationTableName ||'.csv', 'W');  
           UTL_FILE.PUT_Line(v_file, recordsfetched);
           UTL_FILE.FCLOSE(v_file);

 End Loop;

 Close TableName;

End;

当我执行上述过程时出现以下错误

Error report -
ORA-00932: inconsistent datatypes: expected - got -
ORA-06512: at "GENERATECSV", line 47
ORA-06512: at line 1
00932. 00000 -  "inconsistent datatypes: expected %s got %s"
*Cause:    
*Action:

请进一步帮助我。

谢谢。

Answer 1

您正在将两个值（destColumns和destMeasures）选择到一个变量（recordsfetched）中。

rabbitmqctl eval 'Q = {resource, <<"/">>, queue, <<"sharding: sharded_queue - rabbit@hostname - 0">>}, rabbit_amqqueue:internal_delete(Q).'

应该是这样的：

  q2 := 'SELECT ' || destColumns || ',' || destMeasures ||' FROM ' || tablename || '';
  EXECUTE IMMEDIATE  q2 into recordsfetched;

Answer 2

这是一个复杂的过程。但是a不解。为了使第二个动态查询有效，您需要选择一个与组合查询的投影相匹配的记录集。因此，您需要生成一些动态代码来做到这一点。

或者您可以尝试以下操作：

....
Execute Immediate q1;

destColumns := replace( destColumns, ',', q'[||','||]');
destMeasures := replace( destMeasures, ',', q'[||','||]');

q2 := 'SELECT ' || destColumns || ',' || destMeasures ||' FROM ' || tablename || '';
dbms_output.put_line(q2);
EXECUTE IMMEDIATE  q2 into recordsfetched;
....

这会将投影变成单个宽列，可以将其写入文件。

在不知道表名称的情况下将变量声明为行类型时需要帮助

2 个答案: