Prev: Barchart with Graph Template Language
Next: Reshaping the data to run the simple correlation analysis
From: Nathaniel Wooding on 20 Nov 2009 15:53
I have been working today on a data file that dates back to the 1970s and h=
as spent much of its life on a mainframe=2E I am in the process of preparin=
g it for migration to an IT supported database and am cleaning up some glit=
ches that have crept in over the years=2E In particular, I have been focus=
ed on several ids with the values 92=2E2 , 92=2E3 and 92=2E4=2E I was usin=
g code similar to=0D=0A=0D=0AIf station in(92=2E2 , 92=2E3, 92=2E4) then do=
;=0D=0A=0D=0AEtc=0D=0A=0D=0ATo make some corrections to some of the obs wh=
ere a particular variable was not coded uniformly=2E Out of about 400 recor=
ds in this group, all but two behaved themselves and were fixed=2E After a =
couple hours double checking my logic and trying to trap the reason that th=
ese two obs were not behaving themselves, I finally added the statement=0D=
=0A=0D=0AStation =3D Round( station , =2E1);=0D=0A=0D=0AAnd my program star=
ted working! So, I had decimals that were not stored as they appeared ( 92=
=2E3999999999 is, to a human, 92=2E4, etc )=2EWere I to start this particu=
lar data set again, I would most likely avoid the decimal issue but way bac=
k when I first set up the coding scheme, I never anticipated the need to ha=
ve as many ids as we later came to need=2E In all my years of working with =
these data, I have never stumbled on a problem of the decimal representatio=
n not behaving itself=2E I suppose that something may have happened to thes=
e two obs when I used Proc Download to bring the data to my pc so I'll blam=
e the lousy day on that=2E=0D=0A=0D=0AIn any case, for those contemplating =
numeric ids with decimals, beware=2E And, for that matter, perhaps characte=
r ids are better after all=2E=0D=0A=0D=0AHave a very good weekend and, for =
those in the US, have a great Thanksgiving next week=2E=0D=0A=0D=0ANat=0D=
=0A=0D=0A=0D=0ANat Wooding=0D=0AEnvironmental Specialist III=0D=0ADominion,=
Environmental Biology=0D=0A4111 Castlewood Rd=0D=0ARichmond, VA 23234=0D=
=0APhone:804-271-5313, Fax: 804-271-2977=0D=0ACel Phone: 804-205-0752=0D=0A=
=0D=0A=0D=0A=0D=0ACONFIDENTIALITY NOTICE: This electronic message contains=
=0Ainformation which may be legally confidential and or privileged and=0Ado=
es not in any case represent a firm ENERGY COMMODITY bid or offer=0Arelatin=
g thereto which binds the sender without an additional=0Aexpress written co=
nfirmation to that effect=2E The information is=0Aintended solely for the =
individual or entity named above and access=0Aby anyone else is unauthorize=
d=2E If you are not the intended=0Arecipient, any disclosure, copying, dis=
tribution, or use of the=0Acontents of this information is prohibited and m=
ay be unlawful=2E If=0Ayou have received this electronic transmission in e=
rror, please=0Areply immediately to the sender that you have received the m=
essage=0Ain error, and delete it=2E Thank you=2E
 | 
Pages: 1
Prev: Barchart with Graph Template Language
Next: Reshaping the data to run the simple correlation analysis