I guess a cheap way of doing it using the geoDist function suggested by Reeza would be : data haveXY(drop=refLong refLat); retain refLong refLat; set have; if _n_ = 1 then do; refLong = long; refLat = lat; end; id = _n_; x = geodist(lat,refLong,lat,long,"DM"); y = geodist(refLat,long,lat,long,"DM"); run; proc fastclus data=haveXY summary radius=15 strict least=max out=wantClust outseed=seedWant cluster=Site clusterlabel="Site" noprint; var x y; id id; run; The first store coordinates are arbitrarly chosen as a reference. The "DM" option means that your coordinates are in degrees and that the distance is returned in miles. PG ... View more

You can, if you translate your lat/long data to UTM coordinates (say X, Y). I did this recently to group samples taken within 50 meters from each other, it looked like: proc fastclus data=chloro summary radius=50 strict least=max out=chloroClust outseed=seedClust cluster=Site clusterlabel="Site" noprint; var x y; id id; run; PG ... View more

This query does the trick: it excludes all cases where at least one of the missing ROA values occurs on a first quarter (fqtr=1) . Otherwise, it calculates the standard deviations from all non-missing ROA values. There are 33661 std values returned. data beginnings; input begYear @@; datalines; 2003 2004 2005 2006 2007 2008 ; proc sql; create table Roa2Std as select tic, begYear, std(ROA) as RoaStd from fPath.Roa2, beginnings where fyearq between begYear and begYear+2 group by tic, begYear having sum(ROA is not missing or fqtr ne 1) = 12; PG ... View more

Then, it is even simpler: proc sql; /* Companies with females on the board */ create table femComp as select distinct year, ticker from test.Connected_male where char(dirGender,1)="F"; /* Connecting directors : male directors sitting on the board of more than one company with females on the board */ create table connDir as select distinct C.year, C.iddir from test.Connected_male as C inner join femComp as F on C.year=F.year and C.ticker=F.ticker where C.iddir is not missing and char(C.dirGender,1)="M" group by C.year, C.iddir having count(C.ticker) > 1; /* Proportion of connecting directors in each board, every year */ Create table propConn as select C.ticker, C.year, count(D.iddir) / sum(char(C.dirGender,1)="M") as pctConn format=percent10.1 label="Percent connecting" from test.Connected_male as C left join connDir as D on C.year=D.year and C.iddir=D.iddir where ticker is not missing group by C.ticker, C.year; /* print the first 20 results */ proc print data=propConn(obs=20) label; run; -------------------------------------------------------------------                                                                     Percent                    Obs    Ticker                 YEAR    connecting                      1    A                      2001         0.0%                      2    A                      2002          .                      3    A                      2003         0.0%                      4    A                      2004         0.0%                      5    A                      2005         0.0%                      6    A                      2006         0.0%                      7    A                      2007         0.0%                      8    A                      2008         0.0%                      9    A                      2009        86.7%                     10    A                      2010        88.9%                     11    AA                     2001         0.0%                     12    AA                     2002         0.0%                     13    AA                     2003         0.0%                     14    AA                     2004         0.0%                     15    AA                     2005         0.0%                     16    AA                     2006         0.0%                     17    AA                     2007         0.0%                     18    AA                     2008         0.0%                     19    AA                     2009        81.0%                     20    AA                     2010        82.6% ---------------------------------------------------------- PG ... View more

mspak, assuming you can fix the missing iddir problem and that the abrupt jump in the number of companies in 2009 doesn't pose a problem, the following queries should proove useful: proc sql; /* Companies with females on the board */ create table femComp as select distinct year, ticker from Connected_male where char(dirGender,1)="F"; /* Connecting directors : male directors sitting on the board of more than one company with females on the board */ create table connDir as select distinct C.year, C.iddir from Connected_male as C inner join femComp as F on C.year=F.year and C.ticker=F.ticker where C.iddir is not missing and char(C.dirGender,1)="M" group by C.year, C.iddir having count(C.ticker) > 1; /* Total count of connecting male directors per year */ create table connDirCount as select year, count(iddir) as nConn from connDir group by year; /* Total count of male directors per year*/ create table dirCount as select year, count(distinct iddir) as nDir from Connected_male where char(dirGender,1)="M" group by year; /* Proportion of connecting directors to total male director count, every year */ select T.year, C.nConn label="Connecting male directors", T.nDir label="Total male directors", coalesce(C.nConn, 0)/T.nDir as connProp format=percent10.1 label="Percent connecting" from dirCount as T left join connDirCount as C on T.year=C.year; PG ... View more

Your query is logically equivalent to : PROC SQL; CREATE VIEW SASLIB.OUTTAB AS (SELECT FILEA.ALLF1, FILEB.ALLF2 FROM SASLIB.FILEB INNER JOIN SASLIB.FILEA ON FILEA.POL=FILEB.POL ) UNION (SELECT FILEA.ALLF1, FILEB.ALLF2 FROM SASLIB.FILEB INNER JOIN SASLIB.FILEA ON FILEA.LNAME=FILEB.LNAME AND FILEA.FNAME=FILEB.FNAME AND FILEA.DOB=FILEB.DOB); QUIT; which might work better since it avoids the dreaded OR operator. The UNION operation includes unique records only once, even if they appear in both sets. Worth a try. PG ... View more

Look at PROC LOESS. I think it does all that you want to do and a lot more... PG ... View more