Using the "data _null_;" tells SAS that no dataset is to be created. Create the dataset ("data mydataset;) then you use "proc print data=mydataset;" to print it. ... View more

Try this filename in1 FTP "FULLY.QUALIFIED.DATASETNAME.ON.THE.MAINFRAME" host='MainframeNameGoesHere' user="YourUserid" pass="YourPassword"; ... View more

A data step sounds like what you need. If you have SAS Access for the database in question that lets you define a libname statement for the database then this might help. *---> nodupkey option makes "commonvariable" a unique key for each dataset; *---> this is important because none unique records in a merge cause interesting results...; proc sort data=dblib.dbtablename out=dbtable nodupkey; by commonvariable; run; proc sort data=sasdatasetlibname.sasdatasetname out=sasdataset nodupkey; by commonvariable; run; data inboth in_A_only in_B_only; merge dbtable (in=a) sasdataset (in=b); by commonvariable; if a & b then output inboth; else if a then output in_A_only; else if b then output in_B_only; run; The "(in=a)" after the dataset cause a variable name "a" to have value of '1' when a record is read from that dataset and a zero when there is not matching record for the "commonvariable". as far as keeping certain fields... data inboth in_A_only in_B_only; merge dblib.dbtablename (in=a keep=var1 var2 var3) sasdatasetlibname.sasdatasetname (in=b keep=var4 var5 var6); by commonvariable; if a & b then output inboth; else if a then output in_A_only; else if b then output in_B_only; run; All datasets will have var1 var2 var3 var4 var5 var6 in them. You can use a drop or keep statement on the output datasets if they only need some of the variables. Kepp in mind that the dataset "in+A+only" will have values in the var4 var5 var6 variables because the first dataset did not provide any values (no matching record for "commonvariable"). converting the "if a & b" and "else if" into "do-end" constructs can provide youthe opportunity to adjust the values you are interested in. Hope this helps...JJ ... View more

Sometimes trailing blanks are troublesome so try using TRIM. concat = trim(var1) || var2; *---> trim removes trailing blanks; another function you may find of interest...COMPRESS concat = var1 || var2; concat = compress(concat); *---> takes out internal spaces and other chars by default; ... View more

%macro xlsxport(lib,dataset,ssname,sheet); *XPORT a spreadsheet to c:\xxxx\ssname.xls because everyone will have this folder.; PROC EXPORT DATA= &lib..&dataset OUTFILE= "C:\xxxx\yyyy\zzzz\&ssname..xls" DBMS=EXCEL2000 REPLACE; sheet=&sheet; run; %MEND xlsxport; to use this... %xlsxport(work,mydataset,myexcelworkbook,myexcellsheet); You can adjust the path name to fit your needs. ... View more

*---> create format to convert 3 char month to number.; data temp(keep=start label fmtname); retain fmtname "$month2number"; start = 'JAN'; label=1; output; start = 'FEB'; label=2 output; start = 'MAR'; label=3; output; start = 'APR'; label=4; output; ...etc... run; *---> create a temporary format; proc format cntlin=temp; run; data test.datechar1; set test.datechar1; length month $3 year $4; month=substr (date,1,3); year=substr (date,5,4); *---> convert the month to a number.; monthn = put(month,$month2number.); *---> create a SAS date as though it is the first day of the month.; sasdateout = mdy(monthn,1,year); attrib sasdateout length=8 format=EURDFDMY7.; run; ... View more