About TomKari

TomKari · ‎10-30-2019

I'm pleased to hear it! Happy computing, Tom

TomKari · ‎10-29-2019

I think that's what SAS does when it catches an error from the database software, and doesn't know what else to do with it. I think the "Athena Error No: 100, Error Message: No response body." is what SAS is getting back from Athena. Can you look up that error number in the Athena documentation? Are you sure you're successfully connecting to the database using ODBC from the same machine that's running the SAS session? If so, maybe: - credential changes? - firewall / network changes? You could take a look at the SASTRACE system option, using it might return some useful information. Good luck! Tom

TomKari · ‎10-29-2019

640K? Are you kidding?!? When I started, I couldn't submit a job bigger than 256K on our mainframe, one of the biggest in Canada. And don't talk to me about those wimpy personal computers...they hadn't been released yet!

TomKari · ‎10-29-2019

I'm sure you're correct about the DDNAME and PDS member name length limitations. I remember SAS being on the mainframe A LONG TIME before they ported it to other architectures. The variable name limit is funny. Much of SAS is based on PL/I, and it had the COBOL feature of long variable names. FORTRAN at that point was limited to 6 characters. I'm mystified! Tom

TomKari · ‎10-28-2019

Consider using compression. I had a similar situation where I couldn't predict character variable length, and some outliers were VERY long, and when I used the compression option it reduced the overhead to a very manageable amount. Tom

TomKari · ‎10-21-2019

Hi, Michelle First question, are you planning to do this completely within a tool like Enterprise Guide, or are you comfortable with SAS language programming? Next, we really need to get a bit better idea of what you're dealing with. Please show us what your data looks like, and what your desired result looks like. Use text, not a picture, so we can cut and paste it. Tom

TomKari · ‎10-18-2019

On your Windows Start menu you should have a program entry something along the lines of SAS 9.4. Try running that, and if SAS starts, you should see a message in the log, or watch for any other kinds of diagnostics. There may be a problem with SAS starting up. Tom

TomKari · ‎10-18-2019

Boy have I got a task for you! Here's an industrial-strength summarization engine, thanks to PROC MEANS. From Enterprise Guide: Open the dataset you want to summarize (I'm using SASHELP.SHOES as an example). Select Describe -> Summary Statistics On the Data tab, pull the variables you want to summarize into "Classification Variables". I used Region, Product, and Subsidiary. On the Data tab, Pull one numeric variable into "Analysis Variables" (that's the only downside, you need one numeric variable, even though you don't use it for anything.) Make sure it's always non-null. On the "Statistics" "Basic" tab, untick everything except "Number of Observations". On the "Results" tab, untick "Show Statistics", and tick "Save statistics to data set". That will give you a summary dataset. Set "Combinations of classification variables" to "All ways". Click "Run". You'll get a dataset that will show all of the 0-way (summary for the dataset), 1-way (Region, and 2 others), 2-way (Region by Product, and two others) and 3-way (all three variables) summarizations of the data. Tom

TomKari · ‎10-17-2019

I've never tried to use Excel with a different format for thousands and decimal, but I'm wondering is there a way to set Excel to display in that format, and if there is will it fix your problem? Tom

TomKari · ‎10-17-2019

Using one of the localized formats might work. Here's a link to the documentation: https://support.sas.com/documentation/cdl/en/nlsref/61893/HTML/default/viewer.htm#a002604001.htm Tom

TomKari · ‎10-11-2019

If you're dealing with "gnarly" text files, a really good SAS facility to know about is the _infile_ automatic variable. When you input data, it gets set to the contents of a line. If you want to see what your input data looks like to SAS, run the following: Tom data RawText; length InputRecord $32767; infile "C:\data.txt" lrecl=32767; input; InputRecord = _infile_; run;

TomKari · ‎10-03-2019

Two quick comments: In my experience, they are sorted by the order in which they're created. BUT, if you run multiple times, and change the code to add or not create datasets, it will generally sort of keep the datasets that were created on the last run ahead of the datasets that are new to this run. One solution is to delete the datasets from your workspace before running code...then you'll get a nice set in the correct sequence. Once I've tested code, it's rare that I want to see all of the datasets that I've created. So as a last step in the code I delete all of the datasets that I don't need...then the list contains only the remaining datasets. Tom

TomKari · ‎10-03-2019

Yo @Kurt_Bremser! Very well-deserved recognition. It's rare that I check out a discussion, and he's not already there. And, darn it, there's never anything to add to his answers! Looking forward to continuing to learn from Kurt for many, many years! Tom

TomKari · ‎10-02-2019

The advice from @Kurt_Bremser pretty much covers it all. Just one additional comment; sometimes it'll succeed on your MYLIBRARY, and sometimes it'll succeed on WORK, because the balance of free space on both devices will go up and down with the datasets that you and other users are adding and deleting. Completely unpredictable, and exactly what I would expect. Tom

TomKari · ‎10-01-2019

Your logic is a little bit incomplete. It looks like what you want is: The value from the first row for Datum, id, DayFromTxBase, CorrectionBas The last non-missing value for Var1 through Var5 Is this correct?

Online Status	Offline
Date Last Visited	‎07-11-2025 09:14 PM