Hi John, good point, but not a very performance wise solution. Your code implies duplicating columns and rewriting the whole dataset. Astounding solution uses dataset procedure which only reads/writes the descriptor portion of the dataset. I really think that's the best solution you can have. Short code and effective. No need to rewrite the whole data portion of the file when you just want to edit the metadata of the table. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Although having a wrong extension, from what you're saying those files are compressed (zip). Once decompressed you're good to go, no need to open them through anything else. Seems you're just using Notepad++ to rename the extension. Unless Notepad++ is translating something in the process, say line feed/carriage return codes. Either way, I see no need to pass them through Notepad++, SAS will import them fine directly (after decompressed), as Ksharp suggestion. If you care to bypass the unzip process outside SAS, you may try to pipe the unzip command from infile, like this: data _null_; infile 'unzip -p C:\INPUT\MYFILE.bcp' pipe; * -p option pipes the content; input; put _infile_; run; NOTE: -p option of my unzip command line utility pipes the content of the zip, for winzip it's the -c option, you'll have to check the utility your using, and get the correct option flag. Another way, would be to use SASZIPAM engine, which is a "use at your own risk" feature: http://support.sas.com/kb/31/244.html But I wouldn't recommend it for big files like yours. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Hmmm. OK, didn't saw that. You're absolutely right. My mistake, sorry. SUBSTRN won't help there, and it's actually not needed, as I was forgetting the sufficient length of the INPUT var. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Whoops, needs to be modified.  This will generate errors for any names not being truncated.  Add to the WHERE clause: where libname='WORK' and memname='HAVE' and length(name) > 25; Sorry about that . Or simply replace the SUBSTR function with the less used SUBSTRN function which won't generate any errors. More on SUBSTRN here: http://support.sas.com/documentation/cdl/en/lrdict/64316/HTML/default/viewer.htm#a002255112.htm Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

OK, I got it. Code tested was actually the one Ksharp posted, which runs fine. It's the ' char inside the %put It's totally messing the compiler, that's why you get erratic runs. As Ksharp suggested, enclose it inside a masking macro functions so the compiler can get the code right. %str(%') More on masking here: http://support.sas.com/documentation/cdl/en/mcrolref/61885/HTML/default/a001061345.htm Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Runs fine. Nothing wrong about your code. Please share the log and context of your session. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

For a BIG data X TINY data scenario, I guess probably any of those "paths" are a reasonable good solutions. Even a DATA/MERGE would be right if by any means the BIG data happens to be sorted. Performance wise we are probably talking of seconds or milliseconds between each technique. I guess it's more of a practical decision, and taking in mind the whole context of your code. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Tom's answer says pretty much all. I would just add this... As said Macro SAS works mostly at compile time. That's where macro code/vars gets "solved". So that's why there's no interaction with the run time phase, which comes next. I say mostly, because you can actually "fiddle" with Macros at run time. symput, symget, and other datastep macro functions allows you to manipulate Macro vars from the inside of the running datastep. More on this here: http://support.sas.com/documentation/cdl/en/mcrolref/61885/HTML/default/viewer.htm#a003167025.htm Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

The SORT procedure is multithreaded since v9, THREADS option should by set by default, same for CPUCOUNT option to the number of CPUs in your system. Unless your system has no support for multithreading or SAS is under v9, the SORT procedure is already pretty much enabled for parallelism. More on SORT procedure here: http://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#a002473663.htm00    Should you still need to use de MP Connect feature, here's what I use: options autosignon=yes sascmd="!sascmd"; rsubmit process=thread1 wait=no; /*** code here ***/ endrsubmit; rsubmit process=thread2 wait=no; /*** code here ***/ endrsubmit; rsubmit process=thread3 wait=no; /*** code here ***/ endrsubmit; waitfor _all_ thread1 thread2 thread3; options autosignon=no; And a lot more here: http://support.sas.com/rnd/scalability/tricks/index.html Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Try this. %let _MAX=10000; data z; set x; by NAME; drop _:;  * loose temp vars; if first.NAME then _SUM=0;  * reset sum var; _SUM+MONEY;  *sum; if _SUM gt &_MAX then do;  * is it maxed?;    NEW_COL=MONEY-mod(_SUM,&_MAX);    _SUM=&_MAX;  * max sum; end; else NEW_COL=MONEY; run; Cheers from Portugal. ... View more

Hi. Use the dataset option in the object construct, and provide the dataset name you whish to load into the hash. Then loose all the EOFs and until loops. You just need to read an observation from one dataset and match with the other stored in the hash. Be aware that until 9.2, the dataset loaded into the hash should have unique keys. Check the online doc here: http://support.sas.com/documentation/cdl/en/lrdict/62618/HTML/default/a002576871.htm Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

OK forget it. I figured out what you are trying to do. Have you set the connection options previously? [pre] options metaserver="a123.us.company.com" metaport=9999 metauser="metaid" metapass="metapwd" metaprotocol=bridge metarepository="myrepos"; [/pre] Also the documentation states, that you shouldn't have the same alias in the SERVICES file. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more

Hi Chuck. No need of Metadata Server for pipe parallelism, just need an available network port to use and SAS/CONNECT product licensed. The port can be explicitly specified, or you could assign it a name through the SERVICES file. Check this paper: http://support.sas.com/rnd/papers/sugi30/pipeline.pdf Be aware that this technique relies totally on your system multitasking capabilities. Cheers from Portugal. Daniel Santos @ www.cgd.pt ... View more