Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

Re: st: add up variable / quantile

 From Nick Cox To statalist@hsphsun2.harvard.edu Subject Re: st: add up variable / quantile Date Thu, 14 Apr 2011 22:11:18 +0100

```I don't follow what you are doing, but

In any dataset there are just as many observations for one variable as
another. "Observations" means rows, cases, records. Perhaps you mean
by "observations" something else, say  distinct values, non-missing
values, or whatever, but communication is not possible if you don't
explain what you mean so that we understand.

Also, _n - 1 was just an example to show that you can relate values to
those in other observations. It is not a spell to solve your problem,
which unfortunately I am still pretty vague about. It is not an
incantation. You are expected to read up about _n and subscripts to
understand what you are doing.

I think you need to find someone in your University who knows Stata
better, can sit with you, see your data and explain things step by
step.

Statalist unfortunately does not work well when the questioner does
not know quite what to ask precisely and the supposed expert does not
understand the data structure or the exact problem.

Nick

On Thu, Apr 14, 2011 at 9:55 PM, Scharnigg, Stan (Stud. SBE)
<s.scharnigg@student.maastrichtuniversity.nl> wrote:

> Sorry that I keep asking such simple questions, but I don't get the right outcome.
> let's say that var1 has approx. 300.000 observations, and var2 approx 5000 observations.
>
> gen var3 = var1 == var2[_n-1]
> or
> gen var3 = var1[_n-1] == var2
>
> this command gives all the values of var1 a "0", and the rest a "1".
>
> Var1 has 300.000 observations with a value, and approx 2 million with a missing value (.)
> I just don't understand why Stata gives all the values of var1 a "0" and not just the values
> which are the same. I tried many different, because I just want to be certain that I didn't make a
> mistake, but I can't figure this out. I checked the variables and there are lots of matched values between
> them.
>
> I hope you can help me with this.
>
> ________________________________________
> Van: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] namens Scharnigg, Stan (Stud. SBE) [s.scharnigg@student.maastrichtuniversity.nl]
> Verzonden: donderdag 14 april 2011 17:55
> Aan: statalist@hsphsun2.harvard.edu
> Onderwerp: RE: st: add up variable / quantile
>
> aha, then is that the problem. Because var1 has 2.3 million observations and var2 has 5000 observation.
> However, the values of var2 are more than once in var1.
>
> I am gonna try this, thank you :)
> ________________________________________
> Van: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] namens Nick Cox [njcoxstata@gmail.com]
> Verzonden: donderdag 14 april 2011 17:32
> Aan: statalist@hsphsun2.harvard.edu
> Onderwerp: Re: st: add up variable / quantile
>
> gen var3 = var1 == var2
>
> is certainly a good way to create a variable saying whether two
> variables are the same. But what do you mean that "this is not
> working"?
>
> Note that -- necessarily -- this syntax compares values in the same
> observation, and not in different observations.
>
> gen var4 = var1 == var2[_n-1]
>
> illustrates an extension which may be closer to what you seek.
>
> Nick
>
> On Thu, Apr 14, 2011 at 4:04 PM, Scharnigg, Stan (Stud. SBE)
> <s.scharnigg@student.maastrichtuniversity.nl> wrote:
>> I tried a different approach now. I created two new variables (var1, var2), and I want to use those
>> two variables to create a third variable (var3). Var3 needs to be "1" if var1 has the same values as var2
>>
>> So I tried this:
>> gen var3 = 0
>> replace var3 = 1 if var1==var2
>>
>> However, this is not working. Is something like this possible in Stata?
>> ________________________________________
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```