assuming everything is consistent (which might be a big assumption). data want(keep=rec); infile fid; length      txt $100.      rec $100. ; retain rec; input txt & $100.; if substr(txt,1,8)='Customer' then do; /*handles customer1 and customer2*/      rec='';      input txt & $100.; /*name*/      rec=compress(txt);      input txt & $100.; /*address*/      rec=catx(', ',rec,txt);      input txt & $100.; /*city*/      rec=catx(', ',rec,txt);      output;      end; else if substr(txt,1,9)='Amount = ' then do;      rec=substr(txt,10);      output;      end; else if substr(txt,1,7)='Field1:' then do;      rec=substr(txt,9);      output;      end; else if substr(txt,1,7)='Field2:' then do;      rec=substr(txt,9);      input txt & $100.; /*address*/      rec=catx(', ',rec, txt);      input txt & $100.; /*city*/      rec=catx(', ',rec, txt);      output;      end; run; ... View more

hi ... similar to Tom's idea ... data x (drop=temp:); infile datalines truncover; input @1 @'{1:' field1 $4. @1 @'{2:' field2 $5. @1 @'{3:' temp3 $100. @1 @'{4:' temp4 $10. ; if ^missing(temp3) then field3 = substr(temp3,1,find(temp3,'}}')); field4 = input(compress(temp4,'.','kd'),10.); datalines; {1:ABCD}{2:10234}{4:40.45} {1:TRHG}{2:38976}{3:{H132}}{4:567.43} {1:LKMN}{2:23109}{4:10432.67} {1:HGKO}{2:19876}{3:{J987}{F102}}{4:4320.12} {1:DUHN}{2:23098}{3:{G123}{H987}{J234}}{4:345.23} ; run; ... View more