Hi, I have a dataset. Two variables, year and return. What I want to do is for each year I want to count NON-missing observations of return and only keep if number is greater than 10. Now I use a proc sql command, and write something like proc sql;      create table a as      select *      from b      group by year      having count(*) >10      order by year; quit; Will this work? My concern is that count option here will number of observations, no matter there is missing or no missing values in variable return. Thanks. ... View more

Hi, I imported a Stata dataset to SAS. Everything looks correct except for the month variable. I have a month date variable in Stata dataset. It looks like this way, 1991m1, 1991m2, 1991m3,.... When I imported this variable into SAS, It changes to 372, 373, 374.... I wonder how I can correctly convert this variable into a date variable such as 1JAN1991, 1FEB1991,..... Thanks ... View more

Hi, I have a gmm model, and I estimated 16 parameters. I only care about the intercept. Is it possible to output only one parameter in stead of all. I have the following code. ods output parameterestimates=final; proc model data = current;   endo Mean;   exog x1 x2 x3 x4 x11 x12 x13 x14 x15;   instruments _exog_;   parms b0 b1 b2 b3 b4 b11 b12 b13 b14 b15;   Mean=b0+b1*x1+b2*x2+b3*x3+b4*x4+b11*x11+b12*x12+b13*x13+b14*x14+b15*x15;   fit Mean/ gmm kernel=(bart,4,0) vardef=n; run; quit; I only want b0 to be presented. ... View more

I defined 15 macro variables like %let x1=xxx %let x2=xxx %let x3=xxx .... xxx are some variables in my dataset. I wonder how I can call some of them together in some procedures such as proc sql and proc reg. For example, I can write proc reg data=new;      model ret = &x1 &x2 &x3 &x4; run; Can I shorten this code such as model ret = &x1-&x4? I tried this, but not work. Thanks. ... View more

Hi, I have a dataset including independent variables x1-x5, and dependent variable y. I want to do two regressions, y on x1, x2, and x3, and the other regression with y on x4 and x5. Can I use one proc reg command to do two regressions and output to two different files. Something like proc reg data=file outest=model1 model2;      model y=x1 x2 x3;      model y=x4 x5; run; Also, is it easy to output estimates to one data file? Thanks. ... View more

Hi, I have a silly question. I want to full join three tables using proc sql. I tried the following, but doesn't work. I wonder what's wrong with the code. proc sql;      create table new as      select *      from a full join b full join c      on a.id = b.id = c.id; quit; Thanks. ... View more