Thank you art, while this method still need to merge table. ... View more

Thank you! Can we export the graph of gplot to excel file? If yes, how?   Thanks! ... View more

Thanks! ... View more

I would not use something as crude as Windows Notepad for programming work (which inlcudes shell scripting). Actually, I still use vi for scripts - is available on ALL UNIX platforms with the same syntax - works even over a simple telnet connection, as it does not need any function keys/mousing around/etc - can edit files of arbitrary size, although it might change to line mode with sufficiently large files.   If you are working on UNIX from Windows, I recommend using notepad++, as it can write text files in UNIX mode (LF instead of CRLF), provides a hex display mode, and lots of other features that are helpful when writing any type of code. ... View more

SAS already knows how to save the LOG and LST for a program. If you run a program named "test1.sas" then it will automatically create a program named "test1.log" in the current directory to store the LOG and if you print anything to the listing output then it will also create a file name "test1.lst" to store that.  If you don't like those names you can add the -log and -print options to the command line to change them.   If you are running this command from CRON then there is not really much need to add & to the end of the command since it is already running in the background.  If you have two or three SAS programs you want fire up from the SAS script and you want them to run in parallel then you could add the &, but just remember that it means that any later steps you might want to have in there are not then guaranteed to run after the SAS program has finished. ... View more

.bat submit to pc "C:\Program Files\SASHome\SASFoundation\9.3\sas.exe" -SYSIN "C:\Users\test123.sas" -logparm "OPEN=APPEND" -LOG "C:\Users\test123.LOG" .bat submit to mainframe rem ** mainframe aaa submit rem ** submit C:\zschedule rem ** Z: cd "\akaa\asdf\" ftpjob C:\Users\aaaaa\Documents\WORK\Projects\a_batch\a1forthistime >C:\Users\aaaaa\Documents\WORK\Projects\a_batch\a1forthistime.log rem ** rolllog c:\zschedule\a2forthistime.log /a 6 exit .sh sas /home/user/zcrontab/zcrontabtest.sas >/home/user/zcrontab/zcrontabtest.log -logparm "OPEN=APPEND" & ... View more

I agree with both of you! ... View more

%sysjpbid is the id of the job, not the job file.   Thank you. ... View more

With the disclaimer you added .. I agree!   Art, CEO, AnalystFinder.com   ... View more

@Quentin Nothing special about FIND(), it's all functions I reckon: options nofullstimer; %* Create sample and load it in memory; data A(compress=no); A='s'; drop I; do I=1 to 1e8; output; end; run; sasfile A load; %* Select none; data _null_; set A ( where=( put(A,$1.)='X' )) ; run; %* 8.5 seconds; data _null_; set A ; if put(A,$1.)='X' ; run; %* 3.1 seconds; %* Select all; data _null_; set A ( where=( put(A,$1.)='s' )) ; run; %* 20.0 seconds; data _null_; set A ; if put(A,$1.)='s' ; run; %* 3.0 seconds; sasfile A close; WHERE is indeed faster if no function is used *and* most records are not loaded in the PDV: %* Select none; data _null_; set A ; where A>'s' ; run; %* 1.5 seconds; data _null_; set A ; if A>'s' ; run; %* 2.7 seconds; %* Select all; data _null_; set A ; where A>'r' ; run; %* 7.5 seconds; data _null_; set A ; if A>'r' ; run; %* 2.7 seconds; The break-even point here seems to be where about 20% of records get loaded. data A(compress=no); A='s'; do I=1 to 1e8;if I>2e7 then A='a'; output; end; run; sasfile A load; data _null_; set A ; where A>'r' ; run; %* 2.7 seconds; data _null_; set A ; if A>'r' ; run; %* 2.7 seconds; It is very disappointing that functions in a WHERE clause should be slower. There seems to be ample room for optimisation of the software here. ... View more