About BrunoMueller

BrunoMueller · ‎11-27-2014

Well I do not remember whether the Inline formatting was already available with SAS8.2 you might find some user group papers talking about this subject. Have a look here SAS Technical Support Services and Policies you are using a really old version it would make sense to go for a newer version.

BrunoMueller · ‎11-27-2014

The sample code below assumes that the beginning of the two strings is the same. Maybe this will help or at least provide a start. data want; infile cards dlm=","; input var1 : $64. var2 : $64. ; length result $ 64; loopLen = max( length(var1), length(var2) ); * * compare varying length if there is a nonmatch * end loop *; do i = 1 to loopLen; if substr(var1, 1, i) ne substr(var2, 1, i) then do; leave; end; end; * * fill remaining vars *; remain1 = substr(var1, i); remain2 = substr(var2, i); lr1 = lengthn(remain1); lr2 = lengthn(remain2); if lr1 > 1 and lr1 < lr2 then do; result = tranwrd(remain2, strip(remain1), ""); end; else do; result = remain2; end; cards; thebigbad,thebigbadwolf 1REDSTREETMANCHESTER,1REDSTREETGORTONMANCHESTER ab1234,abcöäü1234 abc,abc ;

BrunoMueller · ‎11-27-2014

Have a look at the Inline formatting supported by ODS, SAS(R) 9.4 Output Delivery System: User's Guide, Third Edition Here is some sample code ods escapechar='^'; proc report data=sashelp.class nowd; column age name height weight; define age / order; compute after _page_/ style={ just=l}; length line1 line2 $ 256; line1 = '^{style [font_weight=bold]I want this line font size and need to bold}'; len1 = length(line1); line line1 $varying256. len1; line2 = '^{style [font_size=14pt foreground=cyan]I want this line with different font size and non bold font}'; len2 = length(line2); line line2 $varying256. len2; endcomp; run;

BrunoMueller · ‎11-13-2014

You could use an array for the variables, sort the array and you are set. As with sorting the missing value comes first I changed the order of the variables in the array. See code sample below data have; infile cards dlm="," dsd; input col1 : $12. col2 : $12. col3 : $12. ; array xcol{*} col3 col2 col1 ; putlog "NOTE: Before Sort: " xcol{*}=; call sortc( of xcol{*} ); putlog "NOTE: After Sort: " xcol{*}=; cards; ,beans,rice beets,,lettuce ;

BrunoMueller · ‎10-28-2014

I used SAS9.4M2, where this works just fine. See also Base SAS(R) 9.3 Procedures Guide, Second Edition where it shows how to use a user written function to format values. But only with SAS9.3

BrunoMueller · ‎10-27-2014

You must use the format as seen in my example, like format now db2date.6 now2 datetime25.6; It is important to specify the decimals, the number after the period

BrunoMueller · ‎10-27-2014

Hi You can build your own format using Proc Format, see an example below: proc format; picture db2date (default=26) other = '%Y-%0m-%0d:%0H:%0M:%0s' (datatype=datetime) ; run; data _null_; now = datetime(); now2 = now; format now db2date.6 now2 datetime25.6; putlog (now now2) (=/); run;

BrunoMueller · ‎10-23-2014

you can create your own datetime format to get the what you want. The INTNX function can be used to add or subtract datetime intervals, see sample below: proc format; picture xyyyymmddhh (default=10) low - high = '%Y%m%0d%0H' (datatype=datetime) ; run; data test; now = datetime(); length now_c10 $ 10; now_c10 = put(now, xyyyymmddhh.); now1h = intnx("dthour", now, -1); now_c101h = put(now1h, xyyyymmddhh.); format now datetime22.3; run;

BrunoMueller · ‎10-22-2014

You can make use of arrays for this type of work, see sample code below. For each obs loop through all values and check for the text, if text is found get variable name, compare to a list and add if not found. * test data ; data have; infile cards dlm=","; input subject : $12. F1 : $12. F2 : $12. F3 : $12. F4 : $12. F5 : $12. F6 : $12. ; cards; 10002, 4, 7, 9, 9, 10, 9 10001, 0, 3, 4, Total (JD), 7, Total (PP) 10006, 8, 8, 8, 0, 9, 9 10007, 8, 8, 8, Total (JD), 9, 9 10008, 8, 8, 8, 0, Total (PP), 9 ; * find all variables with some text ; data _null_; set have end=last; array xchar{*} f1 - f6; length dropList $ 1024; retain dropList; drop i; * loop through all vars in array and check for text * if text is found check if we know this variable already * otherwise add it to the list *; do i = 1 to dim( xchar ); if upcase( xchar{i} ) in ("TOTAL (JD)", "TOTAL (PP)") then do; length varname $ 32; drop varname; varname = upcase( vname( xchar{i} ) ); if findw(dropList, varname, ",", "T") = 0 then do; dropList = catx(",", dropList, varname); end; end; end; * replace , by blank and write list into macro var ; if last = 1 then do; dropList = translate(dropList, " ", ","); putlog dropList=; call symputx("dropList", dropList); end; run; * create new DS and drop unwanted vars ; data want; set have; drop &dropList; run;

BrunoMueller · ‎10-20-2014

For the debuuger to work you will need to switch the debugger on, compile the code and then call the application using the AF command with the DEBUG=YES option. This should allow you to call the application that calls your SCL code, once your code is called you should be in the debugger window. You can always write a simple SCL program to call your SCL using the proper parms and then you can see the variables of the ENTRY statement

BrunoMueller · ‎10-14-2014

SAS has the LAG function to look backwards as LinusH pointed out. However there is no looking forward function. But you can kind of simulate this using a second SET statement with the POINT= option so that you will read the I+1 observation, you will need to rename the variables from the second SET statement, so that you can compare. Taking your logic here is some sample code. data have; infile cards dlm=","; input Patient_ID : 8. Admit_Date : anydtdtm. Disch_Date : anydtdtm. Facility_Type : $32. ; format admit_date disch_date datetime19.; cards; 12345,02JAN2014:00:00:00.000,05JAN2014:00:00:00.000,Nursing Home 12345,05JAN2014:00:00:00.000,07JAN2014:00:00:00.000,Hospital 23456,02JAN2014:00:00:00.000,06JAN2014:00:00:00.000,Nursing Home 23456,06JAN2014:00:00:00.000,09JAN2014:00:00:00.000,Hospital 23456,08JAN2014:00:00:00.000,11JAN2014:00:00:00.000,Nursing Home 23456,11JAN2014:00:00:00.000,19JAN2014:00:00:00.000,Hospital 23456,19JAN2014:00:00:00.000,23JAN2014:00:00:00.000,Nursing Home 23456,24JAN2014:00:00:00.000,25JAN2014:00:00:00.000,Hospital 23456,25JAN2014:00:00:00.000,30JAN2014:00:00:00.000,Nursing Home 34567,09JAN2014:00:00:00.000,15JAN2014:00:00:00.000,Nursing Home 45678,13JAN2014:00:00:00.000,20JAN2014:00:00:00.000,Nursing Home ; data want; obs1 = 1; do while( obs1 <= nobs); set have nobs=nobs; obs2 = obs1 + 1; set have( rename=( Patient_ID = pId2 Admit_Date = ad2 Disch_Date = dd2 Facility_Type = ft2 ) ) point=obs2 ; If pId2 = Patient_ID and ft2 = "Hospital" and Facility_Type = "Nursing Home" and datepart(ad2) >= datepart(Admit_Date) and datepart(ad2) <= datepart(Disch_Date) Then Flag = 1; Else Flag = 0; output; obs1 + 1; end; drop obs1 obs2 pId2 ad2 dd2 ft2; run;

BrunoMueller · ‎10-10-2014

Hi Have a go at using inline formatting, see example below: ods escapechar="~"; proc tabulate data=sashelp.cars; class origin type; table type, origin ; label type = "Type of the cars~{NEWLINE}as specified by~{NEWLINE}the car manufacturer" origin = "Origin of the cars~{NEWLINE}where most parts are~{NEWLINE}made and assembled" ; run;

BrunoMueller · ‎10-10-2014

Hi You can use the DATA Step loop to fill the array elements. To make it more dynamic you can find out first how many array elements you need. See sample code below: data have; infile cards; input Subj : 8. Score : 8. ; cards; 001 7 001 6 001 5 001 5 001 4 002 8 002 7 002 6 002 6 ; proc sql noprint; select max( maxN ) into :maxN trimmed from ( select subj , count(*) as maxN from have group by subj ) ; quit; %put NOTE: &=maxN; data want; set have; by Subj; array S{&maxN}; retain S1 - S&maxN; if first.Subj then do; call missing(of S{*}); i = 0; end; i + 1; S{i} = Score; if last.Subj then do; output; end; keep subj S1 - S&maxN; run;

BrunoMueller · ‎10-09-2014

Hi Have a look at this discussion it shows how to create your own rounding function, and then package this into a format. For your example, I guess you need to divide the rounded value by 1000.

BrunoMueller · ‎09-26-2014

The defineData and defineKey methods need a char argument. It is also very important to define the variables from the hash object also in the PDV. See an example below: data student_dim; do student_skey = 1 to 10; student_id = put(student_skey, roman.); output; end; run; data test; length student_skey 8 student_id $ 8 ; call missing(student_skey, student_id); if _n_=1 then do; declare hash e(dataset: 'student_dim', ordered: 'yes'); e.definekey('student_skey'); e.definedata('student_skey'); e.definedata('student_id'); e.definedone(); declare hiter ei('e'); end; do while( ei.next() = 0); putlog _all_; end; run;

Online Status	Offline
Date Last Visited	Thursday