Kurt_Bremser Tracker

Re: Does hash join not honor the memsize limitation?

Kurt_Bremser — Wed, 16 Sep 2026 16:35:27 GMT

Hash objects use your available memory, and that is limited by the MEMSIZE option; the first opinion is correct.

When you have hash objects in production jobs that come close to memory limits, you need to constantly monitor the memory usage during such DATA steps (read the logs!) so you avoid unexpected crashes.

Re: Any way to rename a set of variable names?!

Kurt_Bremser — Tue, 18 Aug 2026 09:28:46 GMT

The first question that arises: why do you need to do this?

This dataset looks very much like one of the dreaded "wide" datasets which always cause trouble and are much harder to handle than "long" datasets. This also smells like data hidden in structure.

The only fully automated way to handle this which comes to my mind is a double TRANSPOSE with a modification of the long dataset:

data _temp_;
input x1 x2 x3 x4;
n = _N_;
datalines;
1 2 3 4
;

proc transpose
  data=_temp_
  out=long
;
by n;
var x:;
run;

data long2;
set long;
substr(_name_,1,1) = "y";
run;

proc transpose
  data=long2
  out=want (drop=_name_)
;
by n;
var col1;
id _name_;
run;

Note that, for the second TRANSPOSE to work correctly, you need a variable (or variables) to identify the original observation and used in the BY.

Let us know how you arrive in this situation in the first place.

Re: SAS is tooo slow

Kurt_Bremser — Mon, 10 Aug 2026 11:19:42 GMT

For working, you always use internal disks connected to the main bus via SATA. The external USB connections are good for backup and similar, but not for real heavy load working.

Re: SAS is tooo slow

Kurt_Bremser — Sun, 09 Aug 2026 11:02:32 GMT

Either you have a completely inadequate disk (read: from the 90's), or you have other processes congesting your computer.

Process monitor from Microsoft Sysinternals can be handy.

Re: 403 Forbidden on welcome.oda.sas.com from mainland China

Kurt_Bremser — Fri, 07 Aug 2026 15:27:55 GMT

Ask your government. But don't tell them you use a foreign VPN to circumvent their "security measures".

Re: SAS Support Communities metrics and their trends

Kurt_Bremser — Thu, 06 Aug 2026 13:44:23 GMT

TBH, a large part of my accepted solutions have been rehashes of other solutions/posts, and "translation" of SAS documentation for the novice user. In short, I was a better search engine.

With the advent of AI-supported searches in Google and others, these have taken over that particular role, and so the people coming here for help find answers much easier before they post a new thread asking the same old, same old.

And they will be pointed towards relevant documentation also, including examples. Knowing where (and how) to look for relevant documentation has always been a daunting task, given the sheer size of SAS doc.

Re: Transpose Account Data which has multiple entries into multiple columns.

Kurt_Bremser — Tue, 04 Aug 2026 14:27:11 GMT

Can there be groups with more than 5 observations?

Re: Transpose Account Data which has multiple entries into multiple columns.

Kurt_Bremser — Tue, 04 Aug 2026 10:18:24 GMT

What do you intend to do with this data? Keep in mind that, for most purposes, the original long layout is better than wide.

Re: Will this work to partition a large data set into 10?

Kurt_Bremser — Fri, 31 Jul 2026 15:33:24 GMT

Run it, then look at the log. This will tell you all you need.

Also see Maxims 4 and 2.

Re: Downloading a directory of *.sas files

Kurt_Bremser — Wed, 22 Jul 2026 16:46:52 GMT

This

filename out zip 'Batchfile 7.6.zip';

is your problem.

You did not use a fully qualified path name, so SAS tries to write to its current working directory, which is usually the place where the script file

WorkspaceServer.sh

is located, and where you do not have write permission.

Use a file name that includes the complete path starting at root (like you did for indir) and pointing to your home directory. You can also use the tilde shortcut for this:

filename out zip '~/Batchfile 7.6.zip';

Re: Weird SAS error messages on quoted strings and white spaces

Kurt_Bremser — Thu, 16 Jul 2026 21:40:47 GMT

Something unbalanced (unbalanced quotes, macro definition without a valid %MEND, ...) has caused your SAS session to become unstable. Start a new session and rerun your code, paying attention to the above.

Re: Allow Custom DateTime Display Formats

Kurt_Bremser — Tue, 07 Jul 2026 08:31:37 GMT

One must make such transformations (assignment of formats, extraction of sub-values etc.) when the dataset is prepared for analysis. This is a job for your data engineers.

Re: Display Numeric IDs Without Thousands Separators

Kurt_Bremser — Mon, 06 Jul 2026 11:51:02 GMT

Are these variables stored as character or numeric in the underlying dataset?

Re: Errors in Importing Excel Data from a Single Worksheet

Kurt_Bremser — Sat, 04 Jul 2026 08:27:17 GMT

You tried to use a library reference (libname) as infile in PROC IMPORT, which cannot work.
Either use a worksheet like a SAS dataset in your Excel library, or use the correct physical filename in the DATAFILE= option of the PROC IMPORT statement.

Re: Can someone explain the answer code for this challenge question?

Kurt_Bremser — Fri, 19 Jun 2026 17:27:52 GMT

@sfletc30 wrote:

Hey so just a followup question for number 3. So I understand what the code is saying now. If last.Code=1, then do these things (which includes writing to the output table with the OUTPUT statement). And last.Code=1 only exists four times in the dataset. So that explains why the rows went from around 140 to only 4.

But why do I have to explicitly tell SAS this? Why doesn't the implicit OUTPUT statement at the end of the code create the four rows that the explicit one does?

Because the implicit OUTPUT writes an observation for every iteration of the DATA step, unless the iteration is cut short by a subsetting IF. And the number of iterations is (in your case) determined by the observations read from the input dataset.

Re: SAS Studio repeated measures over time

Kurt_Bremser — Thu, 18 Jun 2026 09:36:12 GMT

Please show us an example of your input data (just enough observations to illustrate your case), and the expected result from it.

Re: Can someone explain the answer code for this challenge question?

Kurt_Bremser — Thu, 18 Jun 2026 09:33:28 GMT

My solution would have been shorter:

data snowforecast;
set winter2015_2016;
retain FirstSnow;
by Code;
if first.Code then FirstSnow=Date;
if last.Code;
LastSnow=Date;
WinterLengthWeeks=intck('week', FirstSnow, LastSnow, 'c');
ProjectedFirstSnow=intnx('year', FirstSnow, 1, 'same');
format FirstSnow LastSnow ProjectedFirstSnow date7.;
drop Snow Date;	
run;

The code will read all observations, but keep only the value of the first observation in a group in variable firstsnow.

Then, it will only proceed further if the last observation in a group has been read. All following statements (including the implicit OUTPUT at the end of each step iteration) are "jumped over" in all other cases.

This particular form of an IF statement is called "Subsetting IF"; if the condition is not met, the next iteration of the DATA step is immediately started.

Re: Transposing Data

Kurt_Bremser — Thu, 18 Jun 2026 09:19:21 GMT

Since you know the names of the target variables, a simple DATA step will do it:

data have;
input customer $ flag $ count;
datalines;
12345 COE 5
;

data want;
set have;
length coe non_coe 8;
coe = 0;
non_coe = 0;
select (flag);
  when ("COE") coe = count;
  when ("NON_COE") non_coe = count;
  otherwise putlog "ERROR: non-expected flag!";
end;
drop flag count;
run;

But do you really want that wide dataset, or is this for reporting purposes? What is your final goal, and what is your initial input data?

I ask this to avoid having an XY problem.

Re: SAS interruption and problems with reinstallation

Kurt_Bremser — Sun, 14 Jun 2026 15:42:29 GMT

This particular message can mean that SAS is not able to create the WORK library/directory.

Re: Capture SQL View Definition for Programmatic Change

Kurt_Bremser — Thu, 11 Jun 2026 09:00:59 GMT

UNIX systems have very nice tools for exactly such mass text change operations, e.g. awk. Just run that on the original source codes of the SQL views. If those source codes are not available, someone fucked up big time (see Maxim 27).