About Tom

Tom · ‎10-09-2011

?? What do you want it to look like ?? With the first version it generates: Obs SEQ nm1 nm2 nm3 code nm4 lx1 type apv result d1 4 4 dexa tito loli 342 dunga DW2312 A MAT F 08/12/2011 5 4 dexa tito loli 342 dunga PARMA . With the second version it generates: Obs SEQ nm1 nm2 nm3 code nm4 lx1 type apv result d1 comment 4 4 dexa tito loli 342 dunga DW2312 A MAT F 08/12/2011 PARMA

Tom · ‎10-08-2011

What does the extra line represent? Is it another record for sequence number 4 ? data read; infile file1 end=eof truncover; length SEQ 8 nm1-nm3 $6 code 8 nm4 $6 ; length lx1 $7 type $6 apv $3 result $3 d1 8; format d1 mmddyy10.; informat d1 date9. ; retain SEQ nm1 nm2 nm3 code nm4; drop check ; input @1 check $3. @; if check ne " " then do; input @1 SEQ nm1-nm3 code nm4 ; end; else do ; input @1 lx1 type apv result d1 ; output; end; run; Or is it an additional field ? Perhaps a comment? data read; infile file1 end=eof truncover; length SEQ 8 nm1-nm3 $6 code 8 nm4 $6 ; length lx1 $7 type $6 apv $3 result $3 d1 8; length comment $50; format d1 mmddyy10.; informat d1 date9. ; retain SEQ nm1 nm2 nm3 code nm4; drop check check2; input @1 check $3. @; if check ne " " then do; input @1 SEQ nm1-nm3 code nm4 ; end; else do ; input @1 lx1 type apv result d1 ; if not eof then do; input @1 check2 $3. @@; if check2 = ' ' then input comment; end; output; end; run;

Tom · ‎10-08-2011

Why do you think that you cannot change it to not use CALL EXECUTE? Change a program like this: data _null_; set members ; call execute(cats('%nstr(%%mymacro)','(',memname,');')); run; to a program like this: filename code temp; data _null_; set members; file code; put '%mymacro(' memname ');' ; run; %inc code / source2 ;

Tom · ‎10-08-2011

This is because SAS is expanding the macro and passing the statements onto the command stack. You can prevent this by using %NRSTR to protect the macro call from executing at that point. Then SAS will push the macro call onto the stack and it will then execute after the current data step has finished. There are also other timing issues that can be solved by using this technique. Here is an example to show the difference in how the log will look. %macro x; * This is a macro with a few lines of code ; Data _null_; x='Just to see what happens when it is invoked from '; y='call execute and see how it looks in the SAS log.'; run; * end of macro x; %mend x; options mprint; data _null_; call execute('%x;'); run; data _null_; call execute('%nrstr(%%x;)'); run;

Tom · ‎10-08-2011

Art - Can you post your code? I did not have trouble reading those characters. data _null_; do filename='file1.dat','file2.dat'; file dummy filevar=filename recfm=f lrecl=1; do i=1 to 24 ; input char $hex2. +1 @; put char $char1. ; end; input; end; stop; cards; FF D8 FF E0 00 10 4A 46 49 46 00 01 01 01 00 60 00 60 00 00 FF E1 5C 1C FF D8 FF E0 00 10 4A 46 49 46 00 01 01 01 00 60 00 60 00 00 FF E1 5C 22 run; data _null_; infile 'file1.dat'; input; put _infile_ $hex48.; infile 'file2.dat'; input; put _infile_ $hex48.; run; FFD8FFE000104A46494600010101006000600000FFE15C1C FFD8FFE000104A46494600010101006000600000FFE15C22 NOTE: 1 record was read from the infile 'file1.dat'. The minimum record length was 24. The maximum record length was 24. NOTE: 1 record was read from the infile 'file2.dat'. The minimum record length was 24. The maximum record length was 24.

Tom · ‎10-08-2011

The real problem is the directory that Windows is using as the current working directory when it starts SAS. If you open a Command Window and move to the directory with your program in it and type the command: > "c:\Program Files\SAS\SAS 9.2\sas" myprogram It will run SAS 9.2 in batch using myprogram.sas as input and writing the log to myprogram.log and output to myprogram.lst.

Tom · ‎10-08-2011

I am not sure that I understand the structure that you are trying to read. But if the date is always on the row where pgid=509281 and elid=1 and the hour is always on the row where pgid=509281 and elid=2 then you can just merge those rows back together so that you can build a datetime value. data have; input cid pgid elid ev $50.; cards; 212673 509279 1 Turn On or Off Water Service Line 212673 509281 1 13 Oct 2011 212673 509281 2 10 run; data want ; merge have(where=(pgid=509281 and elid=1) rename=ev=cdate) have(where=(pgid=509281 and elid=2) rename=ev=chour) ; by cid; timestamp = input(compress(cdate||':'||chour||':00:00'),anydtdtm.); format timestamp datetime.; put cid= timestamp=; run;

Tom · ‎10-07-2011

Why? You could add another variable and include that in the sort. If your sort variable is ID then you could create a 0/1 variable GROUP using this logic. Then sort by group id and all of the missing values (group=1) will come after the non-missing values. data want; set have; group = missing(id); run; proc sort data=want; by group id; run;

Tom · ‎10-07-2011

Art - The period changes which pass will actually attempt to evaluate the macro variable value. So in examples below using 2,3 and 4 leading &'s you see the effect in the symbolgen logic. For 2 we have &&name.&i first tranlating into &name.2 which then translate into <value of &name>2. But for 3 we have &&&name.&i first translating in &<value of &name>2. So when we set name=name we get &name2 and so the value of name2. -Tom 80 %put 2-> &&name.&i; SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable I resolves to 2 SYMBOLGEN: Macro variable NAME resolves to name 2-> name2 81 %put 3-> &&&name.&i; SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable NAME resolves to name SYMBOLGEN: Macro variable I resolves to 2 SYMBOLGEN: Macro variable NAME2 resolves to acct_num 3-> acct_num 82 %put 4-> &&&&name.&i; SYMBOLGEN: && resolves to &. SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable I resolves to 2 SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable NAME resolves to name 4-> name2

Tom · ‎10-07-2011

You did not define a macro variable named NAME. So &name. will generate an error. You can check how the &&'s are being combined by turning on SYMBOLGEN. 64 %let i=2; 65 %let name2=acct_num; 66 %let name=name; 72 %put 4-> &&&&name&i; SYMBOLGEN: && resolves to &. SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable I resolves to 2 SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable NAME2 resolves to acct_num 4-> acct_num 82 %put 4-> &&&&name.&i; SYMBOLGEN: && resolves to &. SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable I resolves to 2 SYMBOLGEN: && resolves to &. SYMBOLGEN: Macro variable NAME resolves to name 4-> name2

Tom · ‎10-07-2011

Sounds more like the macro is not cleaning up. Perhaps it is leaving a file open? Perhaps if you stick end extra semi-colons after the macro call so that SAS knows that it shouldn't look of parameters. Also check the macro source code. Does it include anY statements before the %MACRO or after the %MEND ? If so then those statements will only work the first time you compile the macro. If you want to recompile the macro you can just include it again. %inc hhmacros(unitspec); %monthday; %inc hhmacros(unitspec); %monthnight; %inc hhmacros(unitspec); %monthBMT;

Tom · ‎10-06-2011

If the data is not sorted then you can just decide whether to replace the found value with the new one. data lookup ; if _n_=1 then do; if 0 then set dates; declare Hash H(); H.DefineKey ('number'); H.DefineData ('date'); H.DefineDone (); /* complete hash table definition */ do until(eof); set dates(rename=(date=nextdate)) end=eof; if h.find() then do; date=nextdate; h.add(); end; else if nextdate > date then do; date=nextdate; h.replace(); end; end; end; set trans; if h.find() then date=.; put number= date=; run;

Tom · ‎10-06-2011

I do not have 9.3, but did you try adding '??' to the input function? when (^missing(input(a, ??best32.))) c = a;

Tom · ‎10-05-2011

IF you are using the "enhanced" editor in Windows then you can take advantage of environment variables that it sets. That's a big IF for those of us using SAS on other operating systems or who just prefer the normal program editor. I prefer to use SUBSTR and LENGTH to deal with pulling just the path from the beginning of a string. Using a translate function runs the risk of confusion if some friendly programmer happended to have created a subdirectory that matches your filename. data _null_; p1=sysget('sas_execfilepath'); p2=sysget('sas_execfilename'); dir=substr(p1||' ',1,length(p1)-length(p2)); call system('cd ' || quote(trim(dir))); put 'Running program: ' p1 / 'Switching to directory: ' dir ; run;

Tom · ‎10-05-2011

You should still double check what is causing the underlying issue. If your account numbers have more than 12 significant digits or require leading zeros then you will not be able to convert the scientific notation generated by the BEST12. format back to the original string.

Online Status	Online
Date Last Visited	4m ago