About art297

art297 · ‎11-13-2019

You will fail the question if you provide the response you chose as correct. Carefully compare the test phrase with the output from each suggestion. Art, CEO, AnalystFinder.com

art297 · ‎11-12-2019

Here is one way: data test; length word $20.; length y $80.; x='sas is the leader in analytics'; n=1; do until (missing(scan(x,n))); word=scan(x,n); if not missing(word) then do; if n eq 1 then do; substr(word,1,1) = upcase(substr(word,1,1)); substr(word,3,1) = upcase(substr(word,3,1)); end; else if n eq 3 then do; substr(word,2,1) = upcase(substr(word,2,1)); substr(word,3,1) = upcase(substr(word,3,1)); end; else do i=2 to length(word) by 2; substr(word,i,1) = upcase(substr(word,i,1)); end; put word; y=catx(' ',y,word); n+1; end; end; put x; put y; run; Art, CEO, AnalystFinder.com

art297 · ‎11-11-2019

There are numerous reasons why your macro won't work. E.g., In the macro you try to use two macro variables (&lwork and &sex) that weren't declared. Also, within the macro, you try to assign values directly to at least one of the macro variables by just using an =. There are probably other reasons. Art, CEO, AnalystFinder.com

art297 · ‎11-11-2019

No. Informat tells SAS how to read data. The FORMAT statement tells SAS how to display variable values. Art, CEO, AnalystFinder.com

art297 · ‎11-08-2019

I agree with @Reeza's suggestion but, to account for leading and trailing blanks, I think you will want to add the strip function to the variables you're comparing. E.g., try the following code: data have; input (date1 date2) ($20.); if _n_ eq 3 then date1=' '||trim(date1); if date1 =: date2 then x=1; if trim(date1) =: trim(date2) then y=1; if strip(date1) =: strip(date2) then z=1; cards; 2019-07-08T08:08:11 2019-07-08T08:08 2019-07-08T08:08 2019-07-08T08:08:11 2019-07-08T08:08 2019-07-08T08:08:11 ; run; Art, CEO, AnalystFinder.com

art297 · ‎11-07-2019

My suggestion would be to use the macro a group of us wrote and presented at SAS Global Forum a couple of years ago. One way to do that would be to run the following code: filename ut url 'https://raw.githubusercontent.com/art297/transpose/master/transpose.sas'; %include ut ; %transpose(data=have, out=want, by=ID BID, Guessingrows=1000, var=dx_year dx_code short_description) If you want to see or download the actual macro and/or the paper, you can find it at: https://github.com/art297/transpose Art, CEO, AnalystFinder.com

art297 · ‎11-07-2019

In that case, I think that running the following will achieve what you want: filename ut url 'http://tiny.cc/untranspose_macro'; %include ut ; %untranspose(data=CRSummary, out=want, by=PatientID Hospital AdmissionDate, prefix=transfusion_, suffix=_cr, id=num, var=_time _product _lot_num, var_first=no) data want; set want (drop=num rename=(_time=time _product=product _lot_num=lot_num)); run; It uses a SAS macro that a group of us wrote and presented at last year's SAS Global Forum. If you want to download the macro and/or read about it, you can get the paper, code, powerpoint and tip sheet at: https://github.com/art297/untranspose Art, CEO, AnalystFinder.com

art297 · ‎11-07-2019

Are BID, Dx_year and dx_code all numeric variables, or are some character variables? Art, CEO, AnalystFinder.com

art297 · ‎11-07-2019

Are the variable names all lower case? In your original example, some characters in the variable names were upper case. i.e., unless you indicate otherwise, my proposed solution will expect that the variables (case sensitive) will be in the form: input PatientID Hospital AdmissionDate transfusion_1_time_cr transfusion_1_product_cr transfusion_1_lot_num_cr transfusion_2_time_cr transfusion_2_product_cr transfusion_2_lot_num_cr transfusion_3_time_cr transfusion_3_product_cr transfusion_3_lot_num_cr; etc, thru transfusion_15_time_cr transfusion_15_product_cr transfusion_15_lot_num_cr; Also, what is the wide dataset's actual filename? Art, CEO, AnalystFinder.com

art297 · ‎11-07-2019

Are the other two sets of variables named the same way. i.e. are they: Transfusion1Dtm and Transfusion1Product or Transfusion_1_Dtm and Transfusion_1_Product etc. or something else? Art, CEO, AnalystFinder.com

art297 · ‎11-07-2019

You've already marked this question as answered, but I had a couple of questions that I'd need answered before proposing what I think will be an easier solution. 1. Do you really have 3 variables named Transfusion1Lot#, Transfusion2Lot# and Transfusion3Lot# or are they called something else? 2. Your stated output file doesn't retain the 1s, 2s or 3s that were in the wide file. Do you need to keep them or are they irrelevant? Art, CEO, AnalystFinder.com

art297 · ‎10-31-2019

Like I said, the monotonic() function isn't documented, so users have only guessed at how it works (see: https://communities.sas.com/t5/SAS-Communities-Library/MONOTONIC-function-in-PROC-SQL/ta-p/475752 ). However, yes, I think the group statement would cause the function to work as you want it to. But, as it isn't documented, there are no guarantees. Art, CEO, AnalystFinder.com

art297 · ‎10-30-2019

That\s why I created file NEED .. so that you could see which records had the minimum count for each group.

art297 · ‎10-30-2019

The WANT you indicated in your initial post didn't match the one you would get from the two proc sorts you showed. The following will, but uses an undocumented function, namely monotonic(), so I wouldn't suggest using it if this is production code. Also, while proc sql appears to read data sequentially, by definition there is no guarantee that it will always work that way: data have; input key mode $ time (x y z) ($); cards; 1 int 1 a b c 1 pho 1 d a b 1 pho 2 dd aa bb 2 int 1 c v a 3 pho 1 a b c 3 pho 1 aa bb cc 3 pho 2 aaa bbb ccc ; run; proc sql; create table need as select *, monotonic() as count from have group by key, mode, time ; create table want (drop=count) as select * from need group by key, mode having count=min(count) ; quit; Art, CEO, AnalystFinder.com

art297 · ‎10-28-2019

If the macro worked on your system, post the code that you used to call the macro and let us know what you weren't able to accomplish. Art, CEO, AnalystFinder.com

Online Status	Offline
Date Last Visited	‎11-26-2024 04:45 PM