What is the numeric value for "Moderate"? ... View more

Add some sample data in a working data step along with desired results ... View more

Put in a data step and find out   data users; input id mu idu; datalines; 1105 0 0 1106 0 1 1107 1 0 1108 1 1 1109 . . 1110 . 1 1111 1 . ; run; data newcode; set users; if mu = 1 and idu = 0 then du = 1; *Marijuana use; if mu = 0 and idu = 1 then du = 2; *non marijuana illegal drugs use; if mu = 1 and idu = 1 then du = 3; *both drugs use; if mu = 1 and idu = . then du = 1; *Marijuana use; if mu = . and idu = 1 then du = 2; *non marijuana illegal drugs use; if mu = 0 and idu = 0 then du = 0; *never use; if mu = 0 and idu = . then du = 0; *never use; if mu = . and idu = 0 then du = 0; *never use; label du = Drug Use; run; results in    id mu idu du 1105 0 0 0 1106 0 1 2 1107 1 0 1 1108 1 1 3 1109 . . . 1110 . 1 2 1111 1 . 1   which looks right.    You could also go    data newcode2; set users; du = 0; if mu = 1 then do; if idu =1 then du = 3; else du =1; end; if idu = 1 then do; if mu = 1 then du = 3; else du =2; end; run; just for fun, although I'm not sure it is an improvement.    I would get in the habit of setting a variable like drugUse2 to zero as a default when you create it so every record is populated.          ... View more

Nobody is going to download those files and nobody (certainly not me) is going to understand or undertake your problem as written.    A little bit more explanation, a working data step, an example of desired results, and things you've tried will g a long way toward helping someone help you.   ... View more

If you simply must have an SQL solution use DISTINCT.   Using BallardW's data statement you can do    proc sql; select distinct a.* from example a; quit; to get x y z 1 2 3 2 3 4 2 3 5 3 4 5   ... View more

Yep.     This { item ,item ,item }   doesn't work. Never loads.    I think I do need { item, item, item }   for the code to like and display it.    I've tried every rearrangement of  "key":' dataset_value :$quote. /',   I can think of with no joy.    "key":' dataset_value :$quote.  ,/'  "key":' dataset_value :$quote. ' ,/ "key":' dataset_value :$quote. ' , /n   How do I do it?  Looking at you Tom.    ... View more

Tom,   Your repost of the $quote. thing was what did it.     I manipulated my programtype array data to look like this as a string: ["11", "10", "9", "16", "14", "1"]   then I wrote it out as  "programname":' programname :$quote. /', "programtype":' programtype /',   without the outside quoting.   That gives me the output I want , "programname":"name" , "programtype":["1", "16"]    except that (unlike you, and I know this will probably tick you off - LOL) I HATE with a passion the commas at the beginning of the line  and would really prefer "programname":"name", "programtype":["1", "16"],   I know the arguments, I just like them at the end. I don't know how to move them so i guess I am stuck until I figure that out.    Thanks for tweaking my brain.            ... View more

Yes. It doesn't matter now but I would have wagered on where year='2008'; ... View more

A numeric variable can only have one format.    Don't worry about how it is stored, format in the presentation.  Surely there is a key or indicator to tell you whether the value in rate is a population or a percent.  Perhaps if there are no populations under 100 you could do it that way.  ... View more

Your "want" makes no sense to me.  Are there more decision rules? Why are there no Total_Customer_OSBal balances for customer 2 for Feb, Mar, Apr, and May?    ... View more

  1. That is bad data structure.   2.  Posting  a working data step so people don't have to retype your data or make some data to work with greatly increases your chances of a useful response.       3. "identify those individuals who do have siblings among the original ID's, and create an indicator" So an indicator (new column) that 4037 has siblings in the data.  Doesn't matter whether 4044 is in or out, 4180 is in so 4037 gets an indicator?   4.  In the bigger picture, what are you really trying to do?     ... View more

I'm wondering why if datediff is in fact a calculated date difference why one would store that instead of calculating it when it is needed. ... View more

Proc SQL is not allowed? So this is homework? ... View more

Not sure what you are really trying to accomplish. Do you have 400,000 cases and 5,600,000 controls? Are you wanting to create a dataset of about 1.6 million observations in which 25 percent of the records (400,000) are cases and 75 percent of the records (1.2 million) are controls selected to match the cases? ... View more