I’m sure everybody ran into an error of some sort when using Oracle. Some errors are really obvious, you are trying to select a column that doesn’t exist, assigning a variable that hasn’t been declared, that kind of errors. But there are also errors that you don’t run into that often, maybe it’s even the first time you run into this error. What do you do when you run into a new error, i.e. new to you. You try to find out what the error is and what might be a resolution. Of course there is the Oracle documentation where you can find all the errors and which might help you with the resolution.

Until today I was under the impression that parameters in PL/SQL were unrestricted in size. That is only partially correct. I was playing around with a simple script to swap two variables without introducing a temporary variable (I know, completely useless, but fun to think about a bit).

As written before, I am currently trying out some things on Oracle 11G after a presentation by Lucas and Alex of AMIS. One of the new features they talked about was the virtual columns option in Oracle. This a column based on an expression that doesn’t really exist in the database.

Some time ago there was a question  (in Dutch) on the utPLSQL forum about filling up a table with a lot of dates and some info about those dates, like the weeknumber, the dayname etc. I created code using persistent columns at first, so the code would run on a pre-11G installation, but I was thinking about how this could be done using the new Virtual Column option in 11G.

During my work for AMIS I came across a (to me at least) new feature of the SQL engine. Instead of building an inline view in the from clause of a statement you can build this ‘inline view’ before the SQL statement and use the results in the from  9ifinalclause.

I’ve seen a presentation on the 11G new features from Alex and Lucas at Amis . One of the questions posed there was when you are using the new follows option in triggers, what would happen if you have two triggers following the same one. A bit like the way it’’s done right now.

The other day I came across a very strange problem in the code I am working on. The code raised a correct error, when executed the first time, but ran with no problems the second time. After a lot of debugging (even built a log package, hopefully so generic it will one day make it to my toolbox) I found out that the problem was a cursor that wasn’t closed properly when the error appeared.