The SAS Course I Really Need

For the last three days I've been on the SAS Essentials course. For the girl with a First in Maths from Kings who has joined straight from university, it's been a stressful time. She's had a lot to make sense of. Mind you, the brains that got her that First mean she is making sense of it quickly.

By contrast my grey head knows a fair amount of SQL and programming and I have a ton of muscle memory and preferred tricks about this stuff. The data I use is in SQL databases or spreadsheets, not SAS datasets, and I do all my report formatting, graphs and the like in Excel so it can be read across to Powerpoint. SAS has a lot of advanced analytics that Excel doesn't, but the chance of us ever using it is small - the audience wouldn't understand it. The SAS modules we have at work are the basic stuff - nothing fancy like data mining and time series forecasting. (Forecasting? What's that? I did more forecasting when thirty years ago in British Rail than anyone does here at The Bank.)

SAS isn't really a language - it's a command-line compiler that's grown like Topsy and now has a batch file  editing interface. So it has all sorts of weird notations (putting : in front of format statements when reading data but not when writing it) conventions and behaviours. The one about using an if-then with no clause after the "then" as a loop control device is something I'm sure not to do in future. As for those sodding semi-colons...

What I need is a conversion course. Throughout the three days I've been relating the SAS stuff to what I do elsewhere. Length statements are like dimensioning variables, but not quite as you don't have to Dim numerics. There's an analogue of the SQL select statement in Keep and so on. But it's been three days and we haven't done inner joins, which is presumably on the advanced course. Ah. another £5,000 or however much in the bank for SAS. I'd like to have covered that much in one three-day course.

The IT education industry has not yet caught up with the fact that their audience is split into two: complete newbies and permanent low-level users of anything, and the rest of us who have reasonable-to-serious chops in programming and data-handling in at least one tool-set. The first group need the basics courses, the second lot need to incorporate the new stuff into their existing knowledge. The courses are different.