What?
A quick note on how I got round one this one.

Why?
Often enough, our requirement is that the latest record from another table is associated with the current row, and often enough we get the latest by ordering the dataset of the subquery. In T-SQL and MySQL, this is not so much of an issue.

I get this error when having to use an ORDER BY clause in a subquery within an Oracle 11g environment.

How?
Consider the following:

---

Applies to:
- MS Windows 7 Enterprise


What?
So I can ping the server the Oracle 11g database sits on. I've installed the Oracle 11g client tools and can connect using sqlplus. But if I tnsping any oracle database, I get something along the lines of:

Message 3511 not found; No message file for product=NETWORK, facility=TNSMessage
3512 not found; No message file for product=NETWORK, facility=TNSMessage 3513 n
ot found; No message file for product=NETWORK, facility=TNSMessage 3509 not foun
d; No message file for product=NETWORK, facility=TNS

Why?
We NEED several instances of Oracle installed as we have different products/services and these run off various instances of Oracle 9i, 10g, and 11g. This article was written for those that need to support these kind of systems and not as an invite for suggestions on how my business should be run.


How?
So there are a variety of solutions that will fix this. However we have certain restrictions in my work environment thanks partly to group policies:

---

What?
This article describes a function that will return random data based on a given value. The function intends to determine the data type and return data that is absolutely irrelevant to the original data but the same length and the same type of data.

Why?
The joy of working with the plethora of applications out there result in me using MySQL, Oracle PL/SQL and Transact SQL in my day job. The equivalent functions have to be written up all in the name of "seamless integration"...

How?

---

What?
How to use the randomization features of Oracle PL/SQL.

Why?
I've been tasked with writing a stored procedure that scrambles data for developers working with some of our databases containing sensitive data.

How?

     SELECT DBMS_RANDOM.option1[(option2)] FROM DUAL;

---

Why?
We have a stored procedure which contains two select queries. The first query will retrieve a student ID number where the input parameter is the student's username. The second query will return data using the student ID number found in the first query. The stored procedure compiles successfully and without any warnings.

So What?
The problem is that if the student does not have a username but has an ID number, then the first query returns NO ROWS and then the second query errors and the whole stored procedure fails. The same problem happens when using these as a subquery which has an empty result set.

Furthermore
I am using this stored procedure in a SQL Server Reporting Services (SSRS) environment and don't want to return any rows if there are no matches (not a blank row either) as one of my reports uses a row counter to display a "No data found" message when no rows are returned. This does not affect the above problem as the stored procedure simply fails and discontinues processing the overall report.

How?
Browsing the web for a solution, most suggest using NVL() but this only replaces a NULL value with a string of your choice; and even if you NVL all returned fields, no rows are returned, and not a row of NULL/blank/empty values. Sounds confusing? That's just me, the answer was using two NVL functions which allowed the stored procedure to return no rows without erroring (ie. where rownum/rowcount = 0).

---

Previously
For a basic Oracle function, visit my article Basic Oracle Function Structure. For a more advanced version which uses cursors to work with SSRS, see my article Oracle Stored Procedures in SSRS.

What?
I asked someone for a simple, easy and basic Oracle stored procedure as my MySQL and T-SQL stored procedures work slightly differently and are easier to pull off. After much umming and aah-ing, I have written this article as a note for me to demo a working stored procedure and how to use cursors. I'm told I need to use this for SQL Server Reporting Services (SSRS) which is why I'm going down this route. For an even simpler stored procedure, I'd imagine you just get rid of the cursor.

Why
A standalone query is faster in the eyes of SSRS, however within SSRS we don't have the capability to run standard PL/SQL commands... unless they're run from within a stored procedure. Our aim is to run a small query first to return the ID of the student, and then to use this number in a second query which we're hoping will be quicker than a straightforward table join.

---

What?
I need to refer to how to create a function in Oracle PL/SQL and sites on the net just attempt to overcomplicate everything and have forgotten how it is to be new to Oracle. I need a function in it's simplest form and if I want to torture my successors, I'll complicate the function myself.

How?
Functions are supposed to return a single value, which is all I need in this case. In my example, I need to submit an Active Directory (AD) username and receive a student ID number instead, all with the aim to improve performance on some SSRS queries which accept either an ID number or an AD name as user parameters.

Syntax:

CREATE [OR REPLACE] FUNCTION function_name
    [ (parameter [,parameter]) ]

    RETURN return_datatype

IS | AS
    [declaration_section]

BEGIN
    executable_section

[EXCEPTION
    exception_section]

END [function_name];

---

Just a quick note as to something that worked for me. Taken from the Oracle forums @ https://forums.oracle.com/forums/thread.jspa?messageID=6861756

Situation
I have one column of values that I've suffixed a comma to but I want to display this as a rows in a table. This is for Oracle PL/SQL.

---