Re: st: Replace with missing value

[Nick Cox <njcoxstata@gmail.com>]

I thought your original question was clear; you just asked a different
question.

Before getting to an answer, I recommend _strongly_ against holding
daily dates as integers such as 20000101. This causes at least two
problems: you need to worry about precision and you get hiatuses in
the data such as that between 20001231 and 20010101. This will bite
you, hard, sooner or later, so I recommend that you convert to Stata
daily dates. Then use date functions such as -year()- to work with
years.

I'll rephrase what appears to be your basic question. You are, or
should be, calculating variances

by <id> <year> : egen var = var(<response>)

but you want to overwrite variances with missing if any values
considered are missing. That could be

by <id> <year> : egen nmissing = total(missing(<response>))

replace var = . if nmissing

Nick
njcoxstata@gmail.com


On 12 April 2013 13:10, André Gyllenram <a_gyllenram@hotmail.com> wrote:

> Thank you for your answer. But i realized that i was not very clear in my first post. Sorry about that.
>
> What i want to do is that i want to compute the variance for different periods and stocks. The problem I have is that stata does not consider missing values when computing the variance.
>
> If i write the code:
>
> by isin: egen variance20001230=var(price) if datum>=20000101 & datum< 2000123
> 0
>
> this gives me the 1-year variance for every stock between 200010101 and 20001230. However it does not consider missing values. If i have any missing values in the price -data for any stock in the time period  20000101-20001230 i want to set the variace for that stock equal to missing.
>
> Maybe this problem is easier to solve?

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/