次の方法で共有


Migrating Oracle to SQL Server using SSMA Error O2SS0007 Check constraint condition not parsed

By Bill Ramos and Ankit Matta, Advaiya Inc.

This blog post covers the reason why SQL Server Migration Assistant (SSMA) for Oracle is not able to convert a BINARY_FLOAT column with a constraint for checking a NAN (Not a Number) condition. In Oracle, the BINARY_FLOAT data type allows an application to store number underflow or overflow values. SQL Server generates an error as soon as a number overflow or underflow is computed and thus can’t store the value in a FLOAT data type.

Error O2SS0007 Check constraint condition not parsed

Background

When you try to convert the schema of a table having a constraint condition which is checking for NAN values, SSMA generates an error “Error O2SS0007 Check constraint condition not parsed” becauseSQL Server doesn’t support floating point conditions like NAN.

Possible Remedies

Consider the below example where we have created a table and altered it by adding a constraint of NAN type:

 

CREATE TABLE NANTEST

(

      COL1 BINARY_FLOAT

);

/

ALTER TABLE NANTEST ADD CONSTRAINT NANCHECK1 CHECK (COL1 IS NOT NAN);

When you try to convert the schema of the above table in SSMA, it generates the below error message:

clip_image001

The solution of the above error can be, divided into 2 steps

Step 1: Create a new table in SQL Server [The equivalent of BINARY_FLOAT in SQL Server is FLOAT(53) ] without including the check constraint.

CREATE TABLE NANTEST

(

      COL1 FLOAT(53)

)

 

Step 2: Modify the application code for inserting the values in this table. Create a TRY… CATCH block which would be restricting the users to insert an unsupported value and generate an error message similar to an error generated from a check constraint.

 

DECLARE

      @VAL FLOAT(53)

BEGIN TRY

      SET @VAL = 1/0

      INSERT INTO NANTEST VALUES (@VAL)

END TRY

 

BEGIN CATCH

      Print ('SQL error is ' + convert(varchar(20),Error_Number()) + ' : ' + convert(varchar(100),ERROR_MESSAGE()))

END CATCH;

 

Trying to use the Transact-SQL ISNUMERIC() function in a check constraint for this scenario is not a viable solution in the check constraint because ISNUMERIC() is designed to check for strings that might represent numeric values and not numeric underflow and overflow conditions.

References

For more information, check out the following references:

· Migrating Oracle to SQL Server 2008 White Paper