Posts

Class on PL/SQL Table Functions at the Oracle Dev Gym

Image
http://bit.ly/dg-tf
A table function is a function that can act like a table inside a SELECT statement. The function returns a collection, and the SQL engine converts that collection into rows and columns that can be manipulated with standard SQL operations (joins, unions, where clauses, etc.).

Far and away the most popular post on this blog is an introduction to a series of articles on table functions:














Given that level of interest in a very interesting feature of PL/SQL, I thought it would be a good thing to give you even more resources to learn about table functions.

So I put together a FREE class at the Oracle Dev Gym on PL/SQL table functions. It consists of four modules and gives you a solid grounding in table function fundamentals:



Each modules consists of a video that covers the basics, followed by a LiveSQL tutorial that dives into more of the details, and gives you an opportunity to run and play with the code. We then finish up the module with quizzes to reinforce and deepen …

How to avoid spamming users from your applications

Image
Does your application send out emails? Lots of emails?

Did you ever get that feeling like someone punched you in the stomach when you realize that you mistakenly sent out hundreds or thousands of emails to your users when you didn't mean to?

I have. It's a terrible feeling. And these days, in the age of GDPR, there can be real consequences for invading the privacy of your users. This post explores how to make sure that, at least when you are developing and testing your code, you do not inadvertently spam your users.

The Oracle Dev Gym sends out lots of different kinds of emails to those players who have opted-in for them, such as:

Results of the quiz you just completedConfirmation of sign-up in a classReminder to take our weekly tournament quizzesHourly reports to site admins with any new errors in our logWeekly activity summaries to quizmasters The Dev Gym is an Oracle Application Express app, so we are able to happily and easily take advantage of the APEX_MAIL package, and it…

How many times does my table function execute?

A left correlation join occurs when you pass as an argument to your table function a column value from a table or view referenced to the left in the table clause. This technique is used with XMLTABLE and JSON_TABLE built-in functions, but also applies to your own table functions.

Here's the thing to remember:
The table function will be called for each row in the table/view that is providing the column to the function.  Clearly, this could cause some performance issues, so be sure that is what you want and need to do.

The following code demonstrates this behavior, for both pipelined and non-pipelined functions.

CREATE TABLE things ( thing_id NUMBER, thing_name VARCHAR2 (100) ) / BEGIN INSERT INTO things VALUES (1, 'Thing 1'); INSERT INTO things VALUES (2, 'Thing 2'); COMMIT; END; / CREATE OR REPLACE TYPE numbers_t IS TABLE OF NUMBER / CREATE OR REPLACE FUNCTION more_numbers (id_in IN NUMBER) RETURN numbers_t IS l_numbers numbers_t := …

The SmartDB Resource Center

Image
I put together this blog post for those interested in learning more about the SmartDB (also or formerly known as "ThickDB") architecture and how to apply it in your applications. I will update it as more resources become available.

What is SmartDB?

Bryn Llewellyn, PL/SQL Product Manager, offers this description:

Large software systems must be built from modules. A module hides its implementation behind an interface that exposes its functionality. This is computer science’s most famous principle. For applications that use an Oracle Database, the database is, of course, one of the modules. The implementation details are the tables and the SQL statements that manipulate them. These are hidden behind a PL/SQL interface.

This is the Smart Database paradigm: select, insert, update, delete, merge, commit, and rollback are issued only from database PL/SQL. Developers and end-users of applications built this way are happy with their correctness, maintainability, security, and performa…

Mutating table errors and multi-row inserts

The Oracle Dev GymPL/SQL Challenge quiz played 28 Apr - 4 May explored the interactions between row-level triggers and multi-row inserts, particularly when it comes to mutating table errors. If you didn't happen to take the quiz and already learn its lesson, here goes.

[Note: you can also click on the link above and play the quiz right now, before you read this post!]

Here's the main rule to keep in mind:
A BEFORE INSERT trigger will not cause a mutating table error as long as the triggering INSERT statement is a single row insert (INSERT-VALUES). Let's take a closer look.

I create a table and a trigger on that table:

CREATE TABLE qz_flowers ( fl_num NUMBER, fl_name VARCHAR2 (30) ) / CREATE OR REPLACE TRIGGER qz_flowers_bir BEFORE INSERT ON qz_flowers FOR EACH ROW DECLARE l_count INTEGER; BEGIN SELECT COUNT (*) INTO l_count FROM qz_flowers; DBMS_OUTPUT.PUT_LINE ('Count = ' || l_count); END; /
The trigger queries from the qz_flowers table, …

Error stack function now (12.2) includes backtrace information!

The DBMS_UTILITY has long (since 10.2) offered three functions that are very handy when either tracing execution or logging errors:
FORMAT_CALL_STACK - answering the question "How did I get here?"FORMAT_ERROR_STACK - answering the question "What was the error?" (or a stack of errors, depending on the situation)FORMAT_ERROR_BACKTRACE - answering the question "On what line was my error raised?" Therefore (and prior to 12.2), if you wanted to get the error information + the line number on which the error was raised, you would need to call both of the "*ERROR*" as in:

CREATE OR REPLACE PROCEDURE p3 AUTHID DEFINER IS BEGIN p2; EXCEPTION WHEN OTHERS THEN DBMS_OUTPUT.put_line (DBMS_UTILITY.format_error_stack); DBMS_OUTPUT.put_line (DBMS_UTILITY.format_error_backtrace); RAISE; END;
Of course, in the real world, you would not display the text on the screen. You would write them to a log table via an autonomous transaction pro…

How do I get the attribute of my object type in SQL?

Image
This question found its way into my In Box yesterday:

I have a table with an object type column. I want to way to get the value of an attribute of that object type in my query. But Oracle keeps telling me "ORA-00904: invalid identifier". What am I doing wrong?

Almost certainly what you are doing wrong is forgetting to use a table alias. Yeah, it's that simple.
Don't forget the table alias. Let's take a look.

I create an object type, use that object type as a column in a table, and insert a couple of rows:

CREATE TYPE food_t AS OBJECT ( NAME VARCHAR2 (100) , food_group VARCHAR2 (100) , grown_in VARCHAR2 (100) ) / CREATE TABLE food_table (id number primary key, my_food food_t) / BEGIN INSERT INTO food_table VALUES (1, NEW food_t ('Mutter Paneer', 'Curry', 'India')); INSERT INTO food_table VALUES (2, NEW food_t ('Cantaloupe', 'Fruit', 'Backyard')); COMMIT; END; /
OK, let&#…