Why ? @ballardw  has already given you the excellent solution . If you really want take ' and ' a dlmstr,Here is the one :   data have; input Subject $40.; datalines; Health science and Math Biology Biology and Geology Arts and Computer Science Geology and Health Science Math ; filename x temp; data _null_; file x; set have; put subject; ; data want; infile x dlmstr=' and ' truncover; input subject1 : $100. subject2 : $100. ; run; ... View more

Maxim 4. Fire up your SAS and do a test run. ... View more

Maybe it would help to make physical model you can try yourself. Take a deck of playing cards and divide it into four stacks by SUIT.  So you have a  CLUB, DIAMOND, HEART and SPADE stack.  That is your input dataset HAVE with four variables. Now run this program. data want; set have; drop club ; run; So first you pick up one card from each stack and put them into your hand.  The hand represents the variables that are active while the data step is running (also called the Program Data Vector or the PDV).  Next place the DIAMOND, HEART and SPACE cards into three new stacks that will represent the WANT dataset.  You will have to discard the CLUB card since that variable is being dropped so there is no stack to place it into. Repeat until you run out of cards in the original four stacks.   How would you modify this example to simulate the effect of a DROP= dataset option on the input dataset? How would you modify this example to simulate the effect of a DROP= dataset option on the output dataset? If you want to get tricky how would you modify it to represent the RENAME statement or the RENAME= dataset option on either input or output dataset?  ... View more

Here's the result. What exactly are you stating "Yes, but only for the PG1 source library version not for the copied LOCAL library version which has had its encoding changed." ?         ... View more

Makes no sense at all. Assuming by COLUMN you mean VARIABLE. Variables cannot be sorted.  Datasets are sorted using the values of one or more BY variables.  Did you mean you want change the order of the variables in the dataset?  Since you are making a NEW dataset by combining two datasets can't you just set the order when you make the new dataset? You can rename variables.  But the variables in any given dataset have to be uniquely named.  You cannot have two variables that are both named TOP in the same dataset. I have no idea what you mean by "trimming" a variable.  It makes no sense at all for a numeric variable.  And also no sense for a character variable in SAS.  SAS character variables are fixed length.  If the string being stored is shorter than the length of the variable it is padded with spaces.   What do you mean about removing "to"?  Are you just looking to change any occurance of that two letter string in any character variable?  What do you want to change it into?  To you want to place it with two space characters? one space character?  Nothing?  What if the string to is part of a larger word, like into.  Would you want to convert into to just in?   Show actual input datasets.  The last bit looks like some attempt at psuedo code. Spell out what code you want to actually run (ie using actual SAS syntax).  For example the syntax for the RENAME= dataset option is space delimiter list of pairs of oldname=newname.  So did you want generate something like merge one(rename=(top1=p1 topx=px)) two;     ... View more