Why do you need the variable to have a normal distribution to determine if other varaibles influence it? Two variables can have any distribution and still have influence on eachother. Can you give further explanation? ... View more

Pronabesh your code has a small error. He does not have name1 in data table_2, so remove that from your merge. data table_3; merge table_1 Table_2; by companyname; run; ... View more

Hizzah, a question I can answer! All joking aside, I hope this is correct. I would bet my salary that you are experiencing issues because some of the fields that you're trying to export have CRLF (carriage return life feed) values within the comments. Sas doesn't know how to handle these in comment fields, so if the field also has a comma sas splits it out into new values, and then creates a blank row, exactly what you are seeing. If you can, try to remove the crlf values from your variables before you export it out, and everything should work out fine. something like this will work. newfield=TRANSLATE(oldfield,'','0D0A'x); You can put this into your loop pretty easily, and then try exporting the datafile. If this doesn't work, please ignore the "bet my salary" portion of my response.. If it does, then glad to be of assistance! Brandon ... View more

You could also specify the value "guessingrows=32000". That will make SAS look at the first 32,000 rows of data to guess the type instead of the first 8. If you can't be guaranteed that the text occurs in the first 32000 rows, then I would suggest not using proc import from an excel file (frankly I suggest this anyways. Excel is just an awful data storage program). ... View more

I agree with Hai.kuo.  If you are planning on doing something with the information that specific account ID's are not present in Table B but are in Table A, there are better ways to go about getting that information... Otherwise why are you doing the join to table B? Assuming your query is correct, there isn't much you can do. You can try either of the following, however the second solution has to do with your Network and if your working on a database located far from your location (i've had this problem before, and it made a 55 minutes query turn into a 2 minute query). 1) Change your where clause into your join clause... This won't help much tho. Proc sql;            Connect to oracle               create table x as                 Select * from connection to oracle                   (select cnt.accnt_id,                           cnt.a_cd,                           cnt.r_cd,                           dst.accnt_id                 from table1 cnt                   left outer join                      table2 dst                   on cnt.accnt_id=dst.accnt_id  and cnt.ts >= '18-OCT-2013'); Disconnect from oracle;   Quit; 2) Make sure your database doesn't have to go over a network if going to another part of the country... If it does, that's a huge problem that you will need to fix through a process or structural change! Good luck! Brandon ... View more

you can't specify a = clause if the value that you are trying to compare is a string and you only have a substring in it. Basically, you have If 03OCT2013  = ...16MAY2013,13JUN2013,11JUL2013,08AUG2013,05SEP2013,03OCT2013,31OCT2013,28NOV2013) Which it does not. CHange your code to use a find clause. THerefore %if find(&varb,&vara) > 1 %then %do;... Where varb is your long string and vara is the date you are looking for. However i'd change your entire macro code to just use the numeric values of dates (instead of the dates themselves) and then use an in clause instead... that's just me however ... View more