Click the Extract code button to copy all custom code of procedures and triggers from the MS SQL Server database to files. This step also adds lines to the code containing SET ANSI NULLS OFF, SET QUOTED IDENTIFIER OFF and SET NOCOUNT ON.

• The MSSQL_StoredProcs_UDTs_marked folder now contains the custom procedures.
• The MSSQL_Triggers_UDTs_marked folder contains the custom triggers.
• The MSSQL_Views folder contains custom views.
• The MSSQL_Unknown_ObjectType folder contains the customizations of an unidentified type. This last folder will only be created if such objects are found in the database.

Click the Run button to re-insert the UDTs into the code of procedures and triggers in the files. All code will be copied and will end up in the StoredProcs_MSSQL_UDTs_replaced and Triggers_MSSQL_UDTs_replaced folders.

The SSMA tool creates a 'sysdb' database containing the emulation functions. Three of these functions have already been moved to the Isah database itself. You will therefore have to change the calls to the SSMA emulation functions in such a way that the functions will refer to the database itself. This will prevent any cross-database references.

To do this, replace 's2ss.' by 'dbo.' in the code in the files.

Check the code for the use of other ssma emulation functions as well. If an emulation function is found, create that ssma emulation function in the Isah database as well and modify the call.

Manually check the use of ASCII, since the ASCII function for Unicode only supports values up to 127. Where it is used to test for <ESC><ESC>, it is no longer required. In all other cases, the UNICODE() function can be used. The <ESC><ESC> code to be deleted will look like this:

If ascii(@old_CountryCode) = 27

And ascii(substring(@old_CountryCode,2,1)) = 27

Begin

Select @old_CountryCode = ''

End

• The Char() function should be replaced by the NChar() function.
• The Char type should be replaced by the NChar type.

You can do this by searching for and replacing whole words (replace the whole word 'Char' by the whole word 'NChar'). This operation can be performed automatically. Most editors will not include parentheses in the search for whole words. Replacement of 'Char' by 'NChar' will therefore change both the Char type and the Char function.

The VarChar type should be replaced by the NVarChar type. You can do this by searching for and replacing whole words.

Change all calls of the char_length_varchar function by the char_length_nvarchar function.

It is necessary to modify:

• all stored procedure calls with non-existing parameters
• stored procedure calls in which the same parameter is passed on more than once.

Non-existing parameters must always be deleted from the list. If the same parameter is passed on more than once, the first occurrence must be preserved, as this is the parameter that was used by Sybase.

1. Click the Check parameters button to generate a list of stored procedure calls that need to be corrected. On the Output tab you will see the path to the Excel sheet with the parameter check information.
2. Correct the errors in the files containing the code and perform this step again to check if all errors have been resolved.

This involves the conversion of strings such as 'abc' to N'abc'.

1. Click the Convert strings button to start the automatic modification process. All code in the files (procedures, triggers and views) is scanned and modified where necessary. Errors are written to the file …\<My Documents>\Isah Migrationbox\ <Migrationbox Project>\Output\UnicodeStringReplace_errors.out.

   After a refresh, any errors are displayed on the Error tab. The error message contains the name of the file and the line on which the string literal begins.

2. Open the files for which an error was reported and correct the code where required.

   The output files contain all occurrences of a quote being used within a string, as it could mean that dynamic SQL is generated here. Perhaps some quotes within the string should be preceded by an 'N'. This automatic conversion does not affect the contents of such strings. Example:

   SELECT @Statement = N'SET @Description = ''Country'''

   This should be the following:

   SELECT @Statement = N'SET @Description = N''Country'''

3. Modify the files manually where necessary.

1. Click the Run button to add the customizations to the database.
2. Click the Check import button to check if all customizations are present. Any missing objects and tables with incorrect column definitions are reported. The result set shows missing objects, missing indexes and columns with incorrect column definitions. Do not forget to check the Output tab, since since the results of the import check are reported there.

Note: In the steps above, it is assumed that there were no unclassified objects in MSSQL_Unknown_ObjectType during the migration. You must check for these objects yourself and replace them where necessary.