SQL & PL/SQL
I was at the supermarket the other day waiting my turn at the checkout behind another guy. The checkout-chick (I'm sure there is a more PC term, I just don't know what it is) just finished scanning his groceries and he asked her to wait until his wife returned with a few last-minute items. I've done it before, so steam didn't start coming out of my ears - yet. Fifteen seconds later she ran up - sorry, sorry - and checked out the last few items, so I wasn't really inconvenienced.
As it turned out, their technique was efficient.
I have started compiling a list of Oracle PL/SQL and SQL bad practices, with the intention of producing a comprehensive catalogue of common programming errors, that can be used as a check-list for code reviews or given to junior developers so that they can learn from the mistakes of others.
For each bad practice, I provided a list of symptoms in the code, an explanation why it causes problems and a list of preferred solutions. I have also listed exceptions to the rules, when they exist.
I have been asked several times recently, how to create "dynamic" views such that a view can be created using "variables". Now I see this question asked on the OraFaq message board, so I thought maybe this is worth posting as an article, so here it is. My answer comes in the form of an Email reply to a friend (why bother to rewrite a prefectly good answer). Please feel free to abuse my answer if you think it deficient, but you better be posting your own alternatives if you do so.
In short an Autonomous Transaction lets a job, commit some data to the database as an ON-THE-SIDE event, without committing data in the MAIN-EVENT. Sounds useful and it can be. But the Autonomous Transaction can also be dangerous. An Autonomous Transaction is kind of like a teenage daughter (I have two) "Oh DAD! You just don't understand me". A lack of understanding is a foundation for trouble. Maybe we can’t understand our teenagers, but we can understand Autonomous Transactions a little better.
If you are on Oracle 9i or higher, you have external tables and pipelined table functions available to meet your ETL needs. But in order to utilize these tools in a dynamic ETL environment, you need to design PL/SQL procedures that can support the loading of files dynamically. This article will provide you with one design that has proven to be very robust and scalable.
In Part I we discussed views, stored procedures, and table functions as the three options of returning record sets from the code stored in the database. In this section I will concentrate on table functions, demonstrating several kinds of solutions that utilize this feature.
Having in my practice found many useful applications for table functions I am a big fan of this feature. Unfortunately, I don’t think it always gets the attention it deserves. Oracle documentation does a decent job of explaining how to code table functions, but the use cases they provide can lead one to believe this feature has no universal application and is meant to solve just a few specific kinds of problems.
So here I am writing even more documentation for my current Gig, and thinking once again, why not post it to OraFAQ and get double duty out of the document. So here is a discussion of the WITH clause that comes with the SELECT statement now. It is easy to use, and handy as all get out, but seems many people have not yet noticed it. Hmm... a SELECT statement that does not start with SELECT.
Analytics are magic. But as with most software products; I seem to use only one percent of the features, ninety nine percent of the time. Indeed, having built warehouses and reporting systems for the last eighteen months, I look back and see that I got a lot done with only three Analytics, SUM, LAG, and LEAD. Knowing how intimidating Analytics can look to those who have not used them, I figured I’d show the uninitiated, how to get in through the back door, with a reduced look at the capability. You can do a lot with very little. We are going to discuss just one Analytic, SUM used to create running totals.
My current client has the crew here doing a good deal of data pivoting in migrating data between several Oracle Systems. I figured to turn some heads in a code review by providing a solution for data pivots which is not based on "locally favored" traditional 3GL coding practices but instead employs what is to me the more natural way of doing it for an Oracle database. I was right, heads did turn. Since I now have to document it, I figured why not post it on OraFAQ once and then tell them its online, come here and read it. So this is an example of pivoting in SQL.