@LisaZ1 wrote: Thanks for your suggestion. it solved my problem. Sorry I wasn't being very clear about my needs. When I said 3 lines, I mean the 3 line with different types (green, blue, and red). To be more specific, I want them to be in 3 different kinds of dashed line so that when I print it out using Black&White printer, I'm able to distinguish which one is which. Thank you! It may be a good idea to mark the message that solves your problem. That way someone else searching the forum or getting here from a search engine can tell the question was solved. ... View more

You want an LOCF (last observation carried forward) for the GDP variable.  Using a conditional SET statement will do that.  Once you understand the implications of the conditional SET, it is strikingly simple.   data return; input year month return; datalines; 2011 1 34 2011 2 33 2011 3 88 2011 4 67 2011 5 39 2011 6 45 2011 7 20 2011 8 29 2011 9 39 2011 10 23 2011 11 28 2011 12 38 run; data GDP; input year month gdp; datalines; 2011 1 345 2011 4 289 2011 7 378 2011 10 589 run; data want; set gdp (drop=gdp in=ingdp) return (in=inreturn); by year month; if ingdp=1 then set gdp (keep=gdp); if inreturn; run; The reason this carries forward the GDP value is because the GDP variable is only read in occasionally.  Whenever the observation in hand is from the GDP dataset  (ingdp=1) only then read in the GDP variable.  It will never be automatically reset to missing (all vars from a SET statement are always retained until replaced by the next time that SET statement is executed).  It will change only when a new GDP observation is encountered.   You could carry forward any number of variables from the gdp dataset.  Just specify them in the appropriate DROP= and KEEP= dataset name options.   No need for lags.  No need for RETAIN. ... View more

thank you so much! I added the date variable and it worked perfectly! ... View more

Thank you for you suggestion. It works. It outputs exactly what I want! Thank you! ... View more

Yeah var cost should be a numeric instead of character. I didn't know that $ means character when creating datelines. Thank you! ... View more

Thank you so much!! I'm new to SAS and I worked for these for almost 2 days. ... View more