Constraints and Limitations

Updated on 2023-04-03 GMT+08:00

View PDF

This section describes the constraints and limitations of DSC.

General

DSC is used only for syntax migration and not for data migration.

If the SELECT clause of a subquery contains an aggregate function when the IN or NOT IN operator is converted to EXISTS or NOT EXISTS, errors may occur during script migration.

Teradata

If a case statement containing FORMAT is not enclosed in parentheses, this statement will not be processed.

For example:

       
          case when column1='0' then column1='value' end (FORMAT 'YYYYMMDD')as alias1

In this example, case when column1= "0", column1= "value" end is not enclosed in parentheses and it will not be processed.

If SELECT * and QUALIFY clauses are both used in an input query, the migrated query returns an additional column for the QUALIFY clause.

An example is as follows:

Teradata query

     
        SELECT * FROM dwQErrDtl_mc.C03_CORP_TIME_DPSIT_ACCT
WHERE 1 = 1
AND Data_Dt = CAST( '20150801' AS DATE FORMAT 'YYYYMMDD' )
QUALIFY ROW_NUMBER( ) OVER( PARTITION BY Agt_Num, Agt_Modif_Num ORDER BY NULL ) = 1;

Query after migration

     
        SELECT * FROM (
SELECT *, ROW_NUMBER( ) OVER( PARTITION BY Agt_Num, Agt_Modif_Num ORDER BY NULL ) AS ROW_NUM1
FROM dwQErrDtl_mc.C03_CORP_TIME_DPSIT_ACCT
WHERE 1 = 1
AND Data_Dt = CAST( '20150801' AS DATE )
) Q1
WHERE Q1.ROW_NUM1 = 1;

In the migrated query, the ROW_NUMBER() OVER(PARTITION BY Agt_Num and Agt_Modif_Num ORDER BY NULL) AS ROW_NUM1 column is returned additionally.

Named references to a table in a query cannot be migrated from subqueries or functions.

For example, if the input query contains a table named foo, DSC will not migrate any named references to the table from a subquery (foo.fooid) or when called from a function (getfoo(foo.fooid)).

      
         SELECT * FROM foo
 WHERE foosubid IN (
                     SELECT foosubid
                       FROM getfoo(foo.fooid) z
                      WHERE z.fooid = foo.fooid
                      );

Oracle

ROWID is not processed in the following case:

      
         SELECT empno,ename ,d.deptno,d.rowid  FROM ( SELECT * FROM employees where deptno is not null) e  , dept d WHERE d.deptno = e.deptno;

The aggregate function in Window functions (RANK and ROW_NUMBER) is not available in the column list.

For example:

      
         SELECT
      TIME, Region, ROW_NUMBER ( ) OVER (ORDER BY SUM (profit) DESC) AS rownumber
        , GROUPING (TIME) AS T
        , GROUPING (Region) AS R
    FROM Sales GROUP BY
        CUBE (TIME, Region)
    ORDER BY
        TIME, Region;

The INSERT FIRST and INSERT ALL queries cannot be migrated if an asterisk (*) is specified in the SELECT statement and the VALUES clause is not specified.

DSC supports the ROWNUM condition specified at the end of the WHERE clause.

For example:

       
          SELECT ROWNUM, ename,empid
FROM employees
WHERE empid = 10 AND deptno = 10 AND ROWNUM < 3
ORDER BY ename;

DSC does not support the ROWNUM function in the UPDATE clause.

For example:

      
         UPDATE tableName SET empno = ROWNUM
   where column = "value";

DSC does not support PARTITION BY LIST in the INDEX clause.

For example:

      
         CREATE TABLE sales(acct_no NUMBER(5)                 
     ORGANIZATION INDEX
             INCLUDING acct_no
             OVERFLOW TABLESPACE example
     PARTITION BY LIST (acct_no)
            (PARTITION VALUES (1)
     PARTITION VALUES (DEFAULT)
                   TABLESPACE example);