Yes, I used arrays and the statements were used so many times with missing values. I noticed that using similar statements on smaller datasets didn't generate negative numbers. I agree with you. I think it is an overflow/overload. hm

Mike Rhoads wrote: > Is it possible that you executed these particular statements so many > times, with so many missing values, that SAS hit an arithmetic overflow > in whatever field it is using to count them? > > Mike Rhoads > Westat > RhoadsM1@Westat.com > > -----Original Message----- > From: owner-sas-l@listserv.uga.edu [mailto:owner-sas-l@listserv.uga.edu] > On Behalf Of hmekouar@hotmail.com > Sent: Thursday, January 26, 2006 9:48 AM > To: sas-l@uga.edu > Subject: Missing values were generated (Number of times is negative) > > > Hi all, > After running a code, my log says: > NOTE: Missing values were generated as a result of performing an > operation on missing values. > Each place is given by: (Number of times) at (Line):(Column). > 63 at 487:53 63 at 487:80 63 at 487:87 > -471080320 at 493:75 -471080320 at 493:104 -471080320 at > 493:111 > -356074240 at 499:77 -356074240 at 499:108 -356074240 at > 499:115 > 126 at 506:19 126 at 506:46 126 at 506:53 > 168 at 512:55 168 at 512:84 168 at 512:91 > 42 at 518:57 42 at 518:88 42 at 518:95 > 832872750 at 524:55 832872750 at 524:83 832872750 at > 524:91 > 818983200 at 536:63 818983200 at 536:92 818983200 at > 536:100 > 20 at 541:71 20 at 541:100 20 at 541:108 > 40 at 553:45 40 at 553:74 40 at 553:82 > I don't understand why the number of times is negative. In the code I > used "<" and "<=" to flag a variable. For instance: > do j=1 to dim(A); > if index0+(30*J)<date<=index0+(30*J)+30 then A(J)=1; > end; > I do have missing values for the variables date and/or index0 for many > observations. But I don't know why the number of times is negative. > Is it something to worry about? > Any help on this is appreciated. > Thank you very much. > HM