PL/SQL provides three different types of collections to hold multiple items of data. They all come with different capabilities and limitations and since these are hard to remember, I provide a condensed showcase of the different collection types in one single code example.
I already wrote about how much I like the XML generating capabilities of modern SQL, but most examples come only with one or two levels of depth.
What if you want to create deeply nested XML and JSON directly from SQL?
Recursive CTEs (WITH-clause) is one addition to the SQL standard that makes the language incredibly powerful. Besides other things it can be used to generate data rows “from thin air” – for example a list of quarters by year.
SQL is your friend!
One of my favorite features of (Oracle-)SQL is how easy it is to turn relational data into other data formats like JSON or XML.
But what if you need to add a namespace to some of the nodes in your generated XML?
I struggled with that until I made a realization about the nature of the xlmns part in XML.
One of the probably most unnerving things in databases is dealing with NULLs, especially when comparing data. The necessary logic to deal with NULLs can lead to very verbose code – but if you’re using Oracle, there is a very neat function you can use.
Wrapping some functionality into a nice little PL/SQL package can make your life so much easier.
I demonstrate this with a small utility-package to deal with Version-Strings.
Ever tried to sort Version-Strings with pure SQL – for example when you want to find the highest Flyway version installed in your database?
Let me show you how you can!
My previous solution to split a string into rows by a delimiter, using pure SQL had a couple of small flaws.
Here is the revisited version that also removes a bit of code duplication.
Sometimes we have data that is not split into columns and rows, but SQL can only really thrive on structured data.
But thanks to recursive queries we can split any string into rows – with pure SQL.
It is a common practice to eliminate possible NULL values from your method’s parameters, and in the past I advocated to use subtypes to do this in PL/SQL.
However, it doesn’t work from SQL…