I saw this on an ideas forum today
and whilst most people would take care of this with a source code control system, its also true that people might do several compilations / tests with their PLSQL source before checking it in officially to their source repository.
So the idea has some merit… and maybe we can roll our own without too much fuss. I make no claims that this is a ‘complete’ solution, but it should get you going
It’s interesting that we still see lots of code out “in the wild” that leans itself to taking forever. Now I’m not talking here about poorly performing SQL etc…because (at least theoretically) no matter how bad the SQL is, presumably it will finish one day, even if that day is years into the future. What I’m referring to here is the default mechanism for locking in Oracle.
If you try access a row that is locked by someone else, you will wait. Don’t get me wrong … that’s a very good thing. The alternative – of letting two people change the same data at the same time is equivalent to saying “I dont care about my data at all”.
And how long will we wait ? Well… we’ll wait forever. We will never stop waiting, until that lock has been released.
I don’t know about you … but for me… forever is long time. A really long time.
RMOUG volunteers compile notifications of webinars, meetings, and meetups from the internet and post them here for everyone to use.
The information technology (IT) industry is always evolving and therefore always changing.
I saw a tweet that made me smile a few days ago:
Like many of you, I started out my working life in IT as a computer programmer. Like some of you, I still am a computer programmer from time to time. I don’t think I’ve had a year of my working life when I did not do either some application development programming or some database infrastructure programming. I am constantly writing small (or not so small) SQL or PL/SQL programs to do what I do in Oracle.
Most people already know about the very cool GATHER_PLAN_STATISTICS hint. If not, you can see an example here
But here’s a common cause of confusion when using it in SQL Plus:
You might be thinking that FIRST_ROWS is pretty much the same as FIRST_ROWS_1, since they are both after the first row. However, there are some important distinctions to make here. The traditional ‘FIRST_ROWS’ parameter is not a genuine cost-based option, more a hybrid between the rule and cost.
I’ve been doing a series on Analytics here but if you need to group data into ranges, you don’t necessarily need a fancy analytic.
Notice in all of the examples below that the date string does not match the format mask.
When a date conversion fails, Oracle tries some similar formats to try succeed. This is actually documented as well here
SQL Plus does not appear to be bothered by the direction of slashes, ie “/” or “\” in its scripts
Hence as long as you have a “C:\tmp” folder on your Windows drive (or wherever your SQL Pus binaary is installed/running from), you can do:
spool /tmp/blah @/tmp/my_file.sql
and it will work on Unix and Windows without alteration.
(This also works for SQLcl as well by the way)
People often think when using the SAMPLE clause, that because they are only dealing with a subset of the data, that immediately it should be a much faster operation. And whilst sometimes this is indeed the case, it is not a guarantee.