Agree! ... View more

Yes, you could test for overdispersion by using the Pearson/DF. The second message is an indication that the model did not converge. A suggestion would be to simplify the model and add complexity back from there. ... View more

That works beautifully.  Thanks. ... View more

That did it.  My root problem is that I didn't know that format assignments were just done in datastep without regard to whether they were in a if-then-do loop.   Also, I was trying to use code like   if &X_VAR = "NatInj_Hand" then do;   when instead I should have been using code like   %if &X_VAR = NatInj_Hand %then %do   So that's that.  Although I've done a lot of macro coding, I haven't used %if and %then much so that was my problem, as well as the aforementioned fact that I didn't know that formats were assigned even if they were in a do loop.   Thanks to all for the help. ... View more

Reset the values when done or when you need them "not retained".   Without a real concrete example it is hard to provide an optimal method.   It is not clear whether you want to set values to missing or to a specific list of values. So example code where you are using a temporary array that needs to be "not retained" with a description of the rules for when "not retained" is desired would help a lot.   From the documentation _TEMPORARY_ creates a list of temporary data elements. Range Temporary data elements can be numeric or character. Tips Temporary data elements behave like DATA step variables with these exceptions: They do not have names. Refer to temporary data elements by the array name and dimension. They do not appear in the output data set. You cannot use the special subscript asterisk (*) to refer to all the elements. Temporary data element values are always automatically retained, rather than being reset to missing at the beginning of the next iteration of the DATA step. Arrays of temporary elements are useful when the only purpose for creating an array is to perform a calculation. To preserve the result of the calculation, assign it to a variable. You can improve performance time by using temporary data elements. ... View more

@n6 wrote: I don't know what icon you mean to right-click.  If I right-click my SAS icon before opening SAS I don't see that. I mean a shortcut on your desktop (or in a folder) that looks like the SAS icon shown in the screenshot in my previous post or, alternatively, an item for SAS in the Windows Start menu (if that exists in your Windows version). Are you saying that you don't see the context menu (with entries from "Open" at the top to "Properties" at the bottom) when you click on one of these SAS icons with the right mouse button? If so, I suspect that your system administrator has blocked context menus (entirely or partially). In this case I'd expect the cmd.exe (which would be another way to start SAS with command line options such as -RSASUSER) to be blocked as well. But you might have access to a config (.cfg) file where you could enter that system option.*   * Edit: See SAS Configuration Files for detailed information about those .cfg files. You could try and copy the default config file sasv9.cfg (typically in C:\Program Files\SASHome\SASFoundation\9.4\nls\en) into your Windows user profile folder C:\Users\username\Documents and add the -NORSASUSER option to the other system options in that copy. As you can see in the documentation, it's still possible, though, that an administrator has specified the opposite option, -RSASUSER, in a place where it overrides your settings (in which case you may want to talk to that person). ... View more

If you are using 64-bit SAS then the error you are getting suggests that the Access database you are reading is 32-bit. To get around this either download and install the SAS PC Files Server, or download the 32-bit Microsoft Jet engine. Both are free so the choice is yours. Another possible option would be to save the Access database as 64-bit which your 64-bit Office should be able to do. ... View more

Okay but it seems to me that if we do that then our edit checking is happening in the Log window at the point of bringing it into SAS and that seems very clunky.  In the first study I mentioned, where it is all character data as soon as we bring it in, we look at each value and if it's clean we output it to a clean dataset and if it's not then we output it to an error dataset and figure out why it's not clean and/or query the site.  We can generate error reporter this way.   If we're bringing in a chunk of data one time and that's it then I can see using the Log window like you describe but we have larger, longer studies where we bring in data again and again over time. ... View more

While it was one big character column to begin with what we did was figure out what kind of variable it was going to be (free text, or a free ranging number like weight, or a date, or a numeric variable but one that had a format, like 1, 2, and 3 with a format of Yes, No and Don't Know), etc. and then deal with each of those separately.  And actually it got more complicated than that for various reasons, one of which was that some dates required an exact date, and for other dates the month and year were enough and we could infer the 15th of the month and for yet other dates just the year was good enough and we could infer July 2 as the month and date.   Maybe there is a more efficient way of doing it but at this point there is no turning back since we have it all working (I think) except for the formats.  I think we're at the point where we have all the variables in a nice SAS dataset and the format names are elsewhere, with the same SAS variable names, and we can merge them and at that point we'll have the variables and their formats in the same dataset.  Of course, all variables will have a format of one kind or another but I'm talking about the formats like 1=Yes, 2=No, 3=Don't Know.  The rest of the formats will already be assigned.  If so I may be able to do that with a data _null_.   I've been busy with other stuff so I haven't tried it yet but I just wanted to post this so you know I haven't vanished. ... View more

@n6 wrote: One thing that is vexing me is that if I do "by ID FormSet" in proc transpose it doesn't do a by on all combinations of ID and FormSet but rather concatenates ID and FormSet into one variable.  There's probably a reason for that, but still. You probably did ID statement instead of BY statement.  BY statement is pretty universal in SAS and always does the same thing. ID statement is specific to PROC TRANSPOSE (although some other procs also use it for slightly different purpose). ... View more

@n6  - I've never had much to do with it myself. I think we have it running on one of our SAS servers so I can check it out at work tomorrow.   Edit - From what I understand it needs to run as a Windows service on your PC (see the Services tab in Windows Task Manager). So that means it can only be started with Administrator rights. Please arrange this with an authorised person. ... View more

PC SAS is just short hand to mean running SAS Display Manager on Windows. Instead of running SAS on Unix or Z/OS (or VMS if it still existed) operating system. Or instead of running Enterprise Guide or SAS/Studio or some other interface instead of Display Manager.   As I said before the idea of the autosave feature is where is it going to dump a copy of what is in the editor, in case you crash SAS or exit without saving.  I am not sure any value is added by going to too much work to constantly be pointing that location different places.   The default is to just write to pgm.asv in the current directory.  Which might work ok for those that launch SAS by moving to a particular current directory and then running the SAS command at the command prompt.  (although that can be a problem when working as a team)  But with typical Windows installations of SAS that is not how SAS is normally launched.  Instead you click on some icon on your desktop (or pick an icon from the "START" menu) or a command is selected for you based on the executable program associated with the file you click on.  None of those methods normally will change the current directory to the location where the file you want to edit is located.  Instead the current working directory is just some arbitrary location setup when the ICON was defined that you used to launch SAS.   The whole concept of dumping the contents of the program editor to a file doesn't really make much sense to me as some that uses SAS display manager.  It is not like SAS is just a text editor. SAS Display Manager is an interactive development environment. What I currently have in the program editor window might just be one step of a many step program that I am working on.  It might even just be some code I am playing with to figure out how to write the program I am working on.  Or how to check the outputs of the steps I ran before that might become part of a program. ... View more

I already put PC Files Server on, way back when, so presumably them updating my computer to Window 10 didn't change that.   I'm going to try to use the test code the other person posted and hopefully it'll work.  Yes, if we go long enough I'll have to go directly to SAS support. ... View more

Yes, the middle part of the path is what I meant.  The way my computer is set up at work there is a long path just to get to my folders so even if a program of mine isn't in a subfolder it has a long pathname and thus I can't see it all.   I tried the submenu thing and it looks to be about the same width as the other one. ... View more