As you say, this is all about precision. -search precision- for some
documentation. In your case,
. gen long ddy =
means that you will not get a -float- type, which is the default.
Alternatively, represent your identifiers using string variables, and
manipulate them by concatenation, just +.
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Lektzian,
David
Sent: 29 January 2008 21:20
To: [email protected]
Subject: st: Mathematical precision question
Dear Statalist,
I have a simple problem that I am hoping somebody can help me with. I
am trying to generate a variable that multiplies three other variables
by quantities and then adds them. Specifically, here is what I am
trying to do:
Var1 = 1972
Var2 = 2
Var3 = 20
I want to generate the variable ddy as follows:
.gen ddy = Var1*1000000 + Var2*1000 + Var3
The correct answer, both on my calculator and in excel is: 1972002020
So, what does Stata come up with? Formatting the answer into the 10
digits necessary to view the answer:
.Format ddy %10.0f
Reveals that Stata's answer is:
1972002048
I'm sure this must be something to do with precision, but I don't know
how to fix the problem. When I run this repeatedly, Stata somewhat
INconsistently gets this operation wrong every time. In the first
example Stata was over by 28. In this next example Stata is under by
36.
Var1 Var2 Var3
1973 002 020 = 1973001984 according to Stata when in reality it
equals 1973002020
Here Stata is closer only going over by 7.
Var1 Var2 Var3
2002 002 055 = 2002002048 according to Stata when in reality it
equals 2002002055
I'm sure this problem must have something to do with precision, but I
need these answers to be exact. Can somebody help explain to me how to
get Stata to perform these mathematical operations correctly?
*As a note, I will mention that I am using these variables as ID
variables which is why I want them exact. There might be a better way
to create ID variables, but I'm a bit of a novice with Stata syntax.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
