LinusH Tracker

Re: Need Guidance on SAS DI Jobs Migration to Azure (590 Jobs / 55 Datamarts)

LinusH — Mon, 28 Sep 2026 10:40:03 GMT

Here on SAS Communities you usually get help on how to use SAS, not how to stop using SAS...

Highly doubt there exists supported migration tools. Usually it's the new supplier (which you for some hasn't told us - Azure based could pretty much be anything that runs in cloud) that supplies migration tools, so contact them/their community for support.

Re: How to periodically clear the log ONLY IF none error?!

LinusH — Mon, 07 Sep 2026 06:56:44 GMT

There are a few options you can play with to prevent the log to grow as fast:
NOMPRINT
NOSYMBOLGEN
NOMLOGIC
NOSOURCE
NOSOURCE2
NONOTES

Re: odbc performance mystery--libname access slow, but sending same SQL via pass-through is really f

LinusH — Fri, 26 Jun 2026 06:49:05 GMT

Yes, the short CPU timings tells us something is going on in SQLS, not SAS.

Re: Add white row before a variable value

LinusH — Thu, 25 Jun 2026 14:52:09 GMT

It seems you try to create a report in a SAS dataset/table, which is not optimal.
If you have somewhat structured source data, there are several SAS procedures to create reports, like PRINT and REPORT.
You can even use a data step with PUT statements for high degree of customization.
Please share a sample of your source data and we could suggest a proper tool to use for your report.

Re: odbc performance mystery--libname access slow, but sending same SQL via pass-through is really f

LinusH — Thu, 25 Jun 2026 06:23:10 GMT

Agree, it seems weird.

Don't know what's going on, but adding more logging/tracing might give some hints:

Compare real and CPU time, gives a hint whether the extra time is spent in SAS or SQLS.
msglevel=i, verifies if the join etc is sent to SQLS
You could platy with some of the other SASTRACE options, like ,,,s and ,,t,

If you don't get closer to an understanding, open a track to SAS could be an option.

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

LinusH — Wed, 17 Jun 2026 09:25:32 GMT

1. Since there's an explicit OUTPUT when last.code, you need to retain the value of firstsnow across observations. All variables are reset for each observation, RETAIN prevents that to happen for selected variable.

2. This set of statements makes sense only to execute when you are at the last observation for your by group /code).

3. In a regular data step, there is an implicit OUTPUT at the end - usually at RUN. When introducing an explicit OUTPUT, there is no implicit OUTPUT any more, and you only write an observation at the OUTPUT statement. In this case when the condition last.code is true.

Re: Recurring question: Automatically Adjust Row Height in Excel

LinusH — Thu, 11 Jun 2026 12:37:06 GMT

With some trial and error and AI conversation, I got all lines visible from your example:

data HAVE;
  x = catx( '0A'x , '0000000000', '4445566778','9999999999','9991123456'); 
run;

ods excel file="print.xlsx" options(sheet_name='Test' 
												 flow="Tables" 
												 row_heights="0,0,0,0,0");          
  proc print data=HAVE noobs ;
  	var x / style(column)={width=2.5cm
                           vjust=top 
                           tagattr="wrap:yes"
                       };
  run;
ods excel close;

But as you can guess, the setting of width= needs to be correlate with the actual length of your strings, which probably will make this hard to use in every day applications....?

Re: SAS and S3 File Systems

LinusH — Wed, 10 Jun 2026 11:14:43 GMT

As I understand it you could mount it to your Kubernetes cluster, but maybe that's not what you are looking for?

Re: Using custom Autoexec at startup of SAS EG

LinusH — Tue, 02 Jun 2026 14:02:18 GMT

There are probably many ways of doing this, so it may come down to your requirements.

How many teams are there that need separate autoexec?

First, I don't think you should bypass the existing set up with the central ones.

The different option within EG will distribute the logic from the server. This will be question of governance.

On the server you could create separate compute contexts (e.g. other SASApp's). This a quite a change and feasible if you don't too many groupings and want other separation between you user groups.

Maybe you could have a %include code in your central autoexec (with some kind of logic that pickup users group and select the appropriate .sas file), or %if %then %else sections in your autoexec code.

Re: PROC EXPORT bug? Placement of WHERE and KEEP in dataset options

LinusH — Wed, 27 May 2026 07:08:30 GMT

This looks like some kind of a bug, report it to SAS tech support.

Re: Snowflake Queries

LinusH — Tue, 05 May 2026 13:08:50 GMT

Please provide the query (and log) so we can help analyze it, ideally with using the options in the link from @Tom.

Re: Export a dataset to csv inserting the delimiter at the end of each row

LinusH — Tue, 21 Apr 2026 10:39:36 GMT

Just to fix the output issue, you could use the of _all_ construct with catx:

data _null_;
	set &Lib..&Mem;
	length out $1000;
	out = cats(catx(';', of _all_),';');
	file "&outfile" mod lrecl=32767 encoding="utf-8";
	put out;
run;

Re: Lesson 7: Accessing DBMS Data with SAS/ACCESS

LinusH — Fri, 17 Apr 2026 06:31:20 GMT

It's pretty much in the message, you need to specify on which server your SAS PC Files Server is executing.
This information should be documented for your environment.

Re: Bulk loading

LinusH — Fri, 17 Apr 2026 06:24:56 GMT

I don't have the "Bulk Load" tab in my advanced settings, probably because we are using SAS/ACCESS to SQL Server, not ODBC. I'm on 9.4M8 Linux.
Not sure what you refer to when you say "advanced option in ODBC set up". Do you mean ODBC Manager (Ican't fins an advanced option here)?
I think it should be enough to set it on the libname, the important thing is to generate the BULKLOAD=YES when your LIBNAME statement is being generated.

Re: Daily SAS Problem?

LinusH — Fri, 27 Mar 2026 12:54:32 GMT

There are lot of coding challenges on internet. You can use those even if they aren't SAS specific, just use SAS to solve them.
One example I've couple of years is "Advent of Code".

Re: How to use SUBSTR to create another variable?

LinusH — Wed, 25 Mar 2026 07:56:12 GMT

As @MarkusWeick suggest, use a numerical date formatted column for this.

data news;
   val = '06May1998'd;
   new_val = intnx('YEAR',val, 1, 'SAME');
   format val new_val date9.;
run;

Re: SAS Enterprise Guide Error Message

LinusH — Mon, 23 Mar 2026 10:28:00 GMT

Could this be caused by VARIANT data types?

Check for lengths in all columns, do any of them exceed 32K?

Also, try implicit pass-through (your query should work with the same performance, explicit pass through shouldn't be necessary here).

Re: SCD Type 2 Loader Modification

LinusH — Fri, 20 Mar 2026 07:51:03 GMT

Another (and not ideal) work-around if you wish to use the existing transformation:
- specify %let etls_debug = 1; in your job(s), or perhaps in appserver_autoexec_usermods.
- Have program (macro preferably) in your postcode of your SCD2 transform (or as a separate transformation) that picks up work.etls_close, modify the end date (for records equal today - drop the rest), and re-apply it to you target table. Of course, if you are expecting a big amount of records to be closed out each batch, this is probably not acceptable from a performance standpoint.

Re: SCD Type 2 Loader Modification

LinusH — Wed, 18 Mar 2026 12:50:36 GMT

This is a outside my realm. I found this (old) paper:

https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://support.sas.com/resources/papers/proceedings/proceedings/forum2007/015-2007.pdf&ved=2ahUKEwi63Ma-vamTAxVDExAIHUkvPEkQFnoECB8QAQ&usg=AOvVaw02Ha3_unK5TGPkuKH_Td9S

May contact you SAS representative or open a SAS track to get better/more detailed guidelines.

One caveat, if you build you own java transformation, that will for not be eligible for migration to Viya 4 when that day comes (please verify with SAS).

Re: SCD Type 2 Loader Modification

LinusH — Wed, 18 Mar 2026 09:04:36 GMT

That is one flaw many have discovered.

Unfortunately, you cannot tweak the DI Studio native transformations, unless you a re a skilled Java developer.

What you can do is take the generated code from the standard transformation, and paste into a User Written Transformation.

It will not be as "nice" as the original one, and some adjustments/configuration to make it work.