Why don't you create a permanent format library. It "sounds" like you are running the PROC FORMAT code in each job possibly, via %INC. but the point is you don't need to do it that way. A library would be more efficient. ... View more

You are telling SAS to create the file without record terminators the F in the record format. RECFM=FB F means fixed, no record termination strings. LRECL=240 is logical record length. So the records do not have termination strings as in "regular" windows files. I can't test it but I think if you change RECFM=V then you will get the proper records. ... View more

You don't mention the criteria for determining the LAST 5. So I choose DATE. The IDGROUP option in PROC SUMMARY makes this very easy, assuming that this is what you want. Then you just use CATX to concatenate the 5 new variables. [pre] data sched; input Course_Code:$4. Location :$16. Begin_Date:date. Teacher &$32.; format BEG: date9.; cards; C001 Seattle 23-Oct-00 Hallis, Dr. George C002 Dallas 4-Dec-00 Wickam, Dr. Alice C003 Boston 8-Jan-01 Forest, Mr. Peter C004 Seattle 22-Jan-01 Tally, Ms. Julia C005 Dallas 26-Feb-01 Hallis, Dr. George C006 Boston 2-Apr-01 Berthan, Ms. Judy C001 Dallas 21-May-01 Hallis, Dr. George ;;;; run; proc summary nway; class course:; output out=wide(drop=_:) idgroup(max(Beg:) out[5](teacher)=); Run; data wide; set wide; length teachers $256; teachers = catx('; ', of teacher_:); drop teacher_:; run; proc print; run; [/pre] ... View more

I not sure my example program is going to be the best approach for that. You’re doling out the specs one a time making it some difficult to understand the ins and outs of your process. How about example data and example output, then you should get better help. ... View more

Perhaps... [pre] data have; input a b c; cards; 1 3 4 3 2 1 4 5 6 ;;;; run; data count4; set have end=eof; array x _numeric_; do _n_ = 1 to dim(x); count4 + x[_n_] eq 4; end; if eof then output; keep count4; run; proc print; run; [/pre] ... View more

To create LIBREFS from meta data it might be easier to use the LIBNAME function. You can also easily test the return code. ... View more

This program works with SAS 9.1.3 on Windows. I don't know if it works in version 8. If PIPE does not work with FILEVAR in version 8 you can do the same thing with multiple data steps. ... View more

You may be able to use this program or some variation. It creates a SAS data set of the name of all sheet in the XLS specified by the WORKBOOK fileref. [pre] filename workbook "path to xls of interest"; data work.sheets; length path script filevar command $256; path = pathname('WORKBOOK'); script = catx('\',pathname('WORK'),'SHEETNAMESS.vbs'); filevar = script; /* write the script */ file dummy1 filevar=filevar; put 'Const ' path=:$quote256.; put 'Set objExcel = CreateObject("Excel.Application")'; put 'With objExcel'; put +3 '.Visible = False'; put +3 '.DisplayAlerts = False'; put +3 'Set objWorkbook = .Workbooks.Open(path)'; put +3 'Set colSheets = .Worksheets'; put +3 'For Each objSheet In colSheets'; put +6 'WScript.echo objsheet.name'; put +6 'Next'; put +3 '.Application.Quit'; put +3 'End With'; /* close the script file by opening another, not used */ filevar = catx('\',pathname('WORK'),'DUMMY.vbs'); file dummy1 filevar=filevar; /* look at the script, not necessary but may be useful */ infile dummy2 filevar=script end=eof; do _n_ = 1 by 1 until(eof); input; putlog _n_ z3. +1 _infile_; end; /* call the script */ command = catx(' ','cscript //nologo',quote(strip(script))); infile dummy3 pipe filevar=command end=eof length=l; do until(eof); input sheet $varying256. l; output; putlog _infile_; end; stop; run; proc contents order=varnum; proc print; run; [/pre] ... View more

How about a method that lets SAS do most of the work and provides a way to easily trap data errors. [pre] data test; input id:$3. (race_african_american race_white race_american_indian)($upcase1.); card; 001 .ny 002 y 003 y 008 yy ;;;; run; proc print; run; proc transpose name=race data=test out=eo(where=(col1 eq'Y')); by id; var race_:; run; data eo errors; set eo(drop=col1); by id; race = propcase(translate(substr(race,5),' ','_')); if first.id and last.id then output eo; else output errors; run; proc print data=eo; run; [/pre] ... View more

Based on the code supplied by the OP a RETAIN to reorder the variables is the most logical approach. I did not say "oh just use retain" I gave reasons why it may be preferred to LENGTH or ATTRIB. I've heard other "programmers" say things like your statement regarding this subject, "Avoid RETAIN ...", "RETAIN is evil" and on an on. It just makes me laugh, at how silly it all sounds. To make you happy perhaps a nice SQL view to the desired variable order should be recommended. ... View more

> Once again, as you mentioned, the RETAIN statement > should be avoided at all cost, especially for > something as subtle as variable ordering. Does that sound as ridiculous to others as it does to me? ... View more

Ok then. Just use a LENGTH statement before the MERGE and name all the variables in the desired order and include the "correct" length attribute for each. ... View more

Why can't you use RETAIN? [pre] retain id numcase Title fName lName Address1 Address2 Apt; [/pre] would be placed before the merge statement and order the variables. The advantage over ATTIB or LENGTH is that you don't need to know about type and length. Since your data are already in SAS data sets, type and length are already established and need not be bothered with. ... View more

200 ... View more

What kind of data step messages did you receive? Try increasing the LRECL, On the FILENAME or on the FILE statement. You may find using DSD DLM='09'x FILE statement option easier than putting the TAB as constant text. ... View more