Proc download allows to get datasets from a server. Efficiency will increase if you can get speciefic columns you need. You can do it by defining a view and downloading this view. Sas will download the actual date and not the view definition as you may expect. Because view is a very small file, the operation of making a view compleets in no seconds.

In the example the data from dataset idl_ovk_kpr will be downloaded. Connect with servers should be established as normal and the server map kkcdb1 should be created. Notice that a view is created in the work map of the server.


rsubmit;
proc sql;
create view idl_ovk_kpr as
select CODMRK, CODDVG, CODKPR, DATING, DATEND, NUMOVK, numklt, NUMVNRKPR,Datreg
from kkcdb1.idl_ovk_kpr
/*where dating between &wrmindt and &wrmaxdt or datend between &wrmindt and &wrmaxdt */
;
quit;
proc download data=idl_ovk_kpr out=srclib.idl_ovk_kpr Connectstatus=y;
run;

problem with modify statement

Keywords: sasbase bi server sas modify datastep problem
Abstract: using modify statement in the datastep is risky

When it comes to updating data, SAS datastep with modify strategy has an importan advantage - it preserves index of the modified dataset. So the consistency of your data is guaranteed at all times. However, I came across a big problem. If something happens during the execution (BI server does not response, somebody runs weird outter join and BI server hangs) the modified dataset got corrupted. Basically you loose it. I experienced this already two times in half year during weekly updates of my datamart. This makes modify statement quite risky to perform on the server side. Even stronger, this fact discourages me to use BI server at all because most of the things work faster on the local computer. (given the fact that you have your major datasources not on the BI server and good local hardare configuration).

My photo

This is my picture, it is not a recent one, but it is me!

(See attached file: maxim.JPG)

Even bellen?

Tijdens de lunch pause , Bas en Bert in de coffee hoek.