PROC MEANS: add new descriptive statistics to count the number of values that are negative,zero, or positive, and null counts
Well I plan on doing it myself. I would have someone else in my company do it, but well, er, that would be me
We will do it internally. Some users will do their own, others will have their local IT support do it. Central IT support will create a depot. As an academic institution, we are very decentralized.
Am I supposed to see your (twocanbazza) full name in here...everywhere else shows your nickname?
I voted that we would have SAS do the install - but I was referring to the EBI/EDI side. The installation and configuration of EBI/EDI was much simpler in 9.2 than 9.1.3, and I suspect that 9.3 installation will be even easier, but why worry? Call in the pros!
We are about to perform a 9.3 install, we have many environments (test/stage/development/production) and I my self have helped in performing 2 of the 9.2 installs, and performing all the migration for one of them.
I voted that I would do this my self, but I might change the word "easy" to something more like "eventful"
It's amazing all of the settings that need to be changed in order to get SAS running and performing to the best of its abilities within a certain environment. (Almost have a small professional Journal that I could probably publish to help others out... but hey that's what these discussion boards are for eh!?)
I am a SAS Affiliate Partner specializing in SAS installs, migrations, upgrades, and performance tuning. So I do all of my own work for our internal lab, but also help with pre-installation planning and architecture and post-installation configuration and performance tuning. Since I originally came from IT, I am able to zero in on I/O and getting SAS to perform as fast as possible so your company can get the most return on its investment in SAS software.
I commonly need the number of non-zero values for each variable, this would be great for that.
I made macro to return those variables from datasets which are having missing values equal to total number of observations of datasets...So it will help in returning counts of non-zero,negative values...gr8...
I think it would be much more efficient to create a format, thus:
value sign (min=1)
low-<0 = 'NEG.'
0 = 'ZERO'
0<-HIGH = 'POS.'
._-.Z = 'MISS'
which could then be used in proc means, proc format, proc univariate, and on and on.
you can easily achieve that with a format (like BennettRutledge already mentioned)
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.