Converting SQL Server Stored Procedure to Oracle

Question

Hi there,
I tried to use SwisSQL to convert my SQL Server stored procedure to Orcale without much success.
Is there anyone who could help me out with this?
My SQL Server Stored Procedure is:
CREATE PROCEDURE [dbo].[SP_BackgroundCheckRequest]
     --MASTER Table
     @MASTER_ID int,
     @PERSONNEL_ID int = NULL,
@DATE_OF_BIRTH datetime = NULL,
     @GENDER varchar(1) = NULL,
     @COUNTRY_OF_BIRTH varchar(3) = NULL,
@TOWN_OF_BIRTH varchar(100) = NULL,
     @STATE_OF_BIRTH varchar(50) = NULL,
     @CHECK_CATEGORY varchar(10) = NULL,
     @CHECK_TYPE varchar(10) = NULL,
     @PRIORITY varchar(10) = NULL,
     @PRIORITY_REASON varchar(100) = NULL,
     @SCREENING_LEVEL nchar(1) = NULL,
     @POSITION nchar(50) = NULL,
     @REQUEST_SOURCE varchar(60) = NULL,
     --NAME_TYPE Table
     @NAME_TYPE varchar(5) = NULL,
     @FAMILY_NAME varchar(50) = NULL,
     @FIRST_NAME varchar(50) = NULL,
     @MIDDLE_NAME varchar(50) = NULL,
     @TITLE varchar(10) = NULL,
     @PREFFERED_NAME varchar(40) = NULL,
     @PREVIOUS_NAME varchar(140) = NULL,
     --ADDRESS_TYPE Table
     @ADDRESS_TYPE varchar(10) = NULL,
     @ADDRESS_LINE_1 varchar(30) = NULL,
     @ADDRESS_LINE_2 varchar(30) = NULL,
     @ADDRESS_LINE_3 varchar(30) = NULL,
     @COUNTRY varchar(3) = NULL,
     @SUBURB varchar(50) = NULL,
     @STATE varchar(50) = NULL,
     @POSTCODE varchar(15) = NULL,
     @START_DATE varchar(10) = NULL,
     @END_DATE varchar(10) = NULL,
     --LICENSE_TYPE Table
     @LICENSE_TYPE varchar(5) = NULL,
     @LICENSE_AGENCY varchar(15) = NULL,
     @LICENSE_NUMBER varchar(100) = NULL,
     @LICENSE_SIGHTED varchar(1) = NULL,
     --PHONE_TYPE Table
     @TELEPHONE_TYPE varchar(5) = NULL,
     @TELEPHONE_NUMBER varchar(20) = NULL,
     @MOBILE_TYPE varchar(5) = NULL,
     @MOBILE_NUMBER varchar(20) = NULL,
     @EXTENSION_TYPE varchar(5) = NULL,
     @EXTENSION_NUMBER varchar(20) = NULL,
     --PASSPORT_TYPE Table
     @PASSPORT_TYPE varchar(5) = NULL,
     @PASSPORT_NUMBER varchar(50) = NULL,
     @PASSPORT_COUNTRY varchar(3) = NULL,
     @PASSPORT_SIGHTED varchar(1) = NULL
AS
BEGIN TRANSACTION
INSERT into MASTER (MASTER_ID, PERSONNEL_ID, DATE_OF_BIRTH, GENDER, COUNTRY_OF_BIRTH,
     TOWN_OF_BIRTH, STATE_OF_BIRTH, CHECK_CATEGORY, CHECK_TYPE, PRIORITY, PRIORITY_REASON,
     SCREENING_LEVEL, POSITION, REQUEST_SOURCE)
     VALUES (@MASTER_ID, @PERSONNEL_ID, @DATE_OF_BIRTH, @GENDER, @COUNTRY_OF_BIRTH,
     @TOWN_OF_BIRTH, @STATE_OF_BIRTH, @CHECK_CATEGORY, @CHECK_TYPE, @PRIORITY, @PRIORITY_REASON,
     @SCREENING_LEVEL, @POSITION, @REQUEST_SOURCE)
IF @@ERROR <> 0
BEGIN
     ROLLBACK
     RAISERROR ('Error inserting values into MASTER table!', 16, 1)
     RETURN
END
INSERT into NAME_TYPE (MASTER_ID,NAME_TYPE,FAMILY_NAME,FIRST_NAME,MIDDLE_NAME,TITLE)
     VALUES (@MASTER_ID,@NAME_TYPE,@FAMILY_NAME,@FIRST_NAME,@MIDDLE_NAME,@TITLE)
IF @@ERROR <> 0
BEGIN
     ROLLBACK
     RAISERROR ('Error inserting values into NAME_TYPE table!', 16, 1)
     RETURN
END
INSERT into ADDRESS_TYPE (MASTER_ID,TYPE,ADDRESS_LINE_1,ADDRESS_LINE_2,ADDRESS_LINE_3,
          COUNTRY,SUBURB,STATE,POSTCODE,START_DATE,END_DATE)
     VALUES (@MASTER_ID,@ADDRESS_TYPE,@ADDRESS_LINE_1,@ADDRESS_LINE_2,@ADDRESS_LINE_3,
          @COUNTRY,@SUBURB,@STATE,@POSTCODE,@START_DATE,@END_DATE)
IF @@ERROR <> 0
BEGIN
     ROLLBACK
     RAISERROR ('Error inserting values into ADDRESS_TYPE table!', 16, 1)
     RETURN
END
INSERT into LICENSE_TYPE (MASTER_ID,TYPE,AGENCY,NUMBER,SIGHTED_YN)
     VALUES (@MASTER_ID,@LICENSE_TYPE,@LICENSE_AGENCY,@LICENSE_NUMBER,@LICENSE_SIGHTED)
IF @@ERROR <> 0
BEGIN
     ROLLBACK
     RAISERROR ('Error inserting values into LICENSE_TYPE table!', 16, 1)
     RETURN
END
INSERT into PHONE_TYPE (MASTER_ID,TYPE,NUMBER)
     VALUES (@MASTER_ID,@TELEPHONE_TYPE,@TELEPHONE_NUMBER)
IF @@ERROR <> 0
BEGIN
     ROLLBACK
     RAISERROR ('Error inserting Telephone number into PHONE_TYPE table!', 16, 1)
     RETURN
END
IF ((@MOBILE_TYPE <> NULL) AND (@MOBILE_NUMBER <> NULL))
BEGIN
     INSERT into PHONE_TYPE (MASTER_ID,TYPE,NUMBER)
          VALUES (@MASTER_ID,@MOBILE_TYPE,@MOBILE_NUMBER)
     IF @@ERROR <> 0
     BEGIN
          ROLLBACK
          RAISERROR ('Error inserting Mobile number into PHONE_TYPE table!', 16, 1)
          RETURN
     END
END
IF ((@EXTENSION_TYPE <> NULL) AND (@EXTENSION_NUMBER <> NULL))
BEGIN
     INSERT into PHONE_TYPE (MASTER_ID,TYPE,NUMBER)
          VALUES (@MASTER_ID,@EXTENSION_TYPE,@EXTENSION_NUMBER)
     IF @@ERROR <> 0
     BEGIN
          ROLLBACK
          RAISERROR ('Error inserting Extension number into PHONE_TYPE table!', 16, 1)
          RETURN
     END
END
INSERT into PASSPORT_TYPE (MASTER_ID,NUMBER,COUNTRY,SIGHTED_YN)
     VALUES (@MASTER_ID,@PASSPORT_NUMBER,@PASSPORT_COUNTRY,@PASSPORT_SIGHTED)
IF @@ERROR <> 0
BEGIN
     ROLLBACK
     RAISERROR ('Error inserting values into PASSPORT_TYPE table!', 16, 1)
     RETURN
END
COMMIT

Answer 1

First, a basic concept. That also illustrates how different Oracle is (and PL/SQL) from SQL-Server (and T-SQL).
PL/SQL integrates two different languages. The PL language. The SQL language. It allows you to code SQL source code natively inside the PL language. The PL compiler is clever enough to do the rest - make calls to the SQL engine to create SQL cursors, bind PL variable values to bind variables in the SQL code. Etc.
PL is a "proper" programming language. It is much like Pascal (it is based on Ada, a close family member of Pascal). It is nothing at all like T-SQL.
Okay, now for the very basic rule for Oracle development.
Maximize SQL. This means using the SQL language to crunch data. It is the "closest" to the data. It is designed and optimised for dealing with data. Do your data processing using SQL.
Minimize PL/SQL. This means using the PL language not to crunch data, but to provide the conditional logic. Implement business rules. And then have the SQL language crunch the data as the data processing language.
Attempting to directly translate T-SQL code into PL/SQL is flawed. It is like trying to make coffee with teabags. Yeah, the tea may have caffeine it, but it is not coffee.
To do what that T-SQL script does, insert a row into a table, a typical PL/SQL equivalent will look as follows:
create or replace procedure InsertMaster( masterRow master%rowtype ) authid definer is
.. types and variables declared here
begin
  -- masterRow is a record structure that matches the actual MASTER table
  .. apply business rules and validation to the data.. raising user exceptions as needed
  -- e.g. the POSITION column must be upper case
  masterRow.position := upper(masterRow.position);
  -- after the business logic and data validation we can add the row to the table
  -- (PL will make the INSERT call to the SQL engine)
  insert into master values masterRow;
end;No commit. The caller does the business transaction. It needs to decide when to commit to maintain the integrity of the data in the database. It may need to do several more calls, before committing.
With this approach, we can give the application schema (the caller) execute privs on this procedure. This procedure is defined with definer rights. This means it runs with the privs of owner of that procedure and MASTER table. It is trusted code. It does all the correct stuff to add a valid and checked row to the MASTER table. So we give the app schema execute privs on the procedure and no insert priv on the MASTER table.
The only way the caller can add row to the MASTER table is via this trusted procedure of ours.
And this is a typical approach in Oracle - using the PL/SQL (both PL language and SQL language) code to create an interface to a logical database in Oracle. Abstracting the complexities of SQL from the caller. Moving business and validation into PL. Allowing us the flexibility of modifying business and validation logic, without touching a single byte of caller/client code. Allowing us to manage and tune SQL performance without dealing with (badly designed and coded) SQL from a client - as the SQL resides in PL/SQL packages and procedures and functions.