Calculate mean when values are missing

[Aline Campos]

I need to create a variable that is the mean of the observations from three other variables. I did this:
gen CPI_foodweight_20142016 = (CPI_foodweight_2014+CPI_foodweight_2015+CPI_foodw eight_2016)/3

However, some of the the values are missing for one or two of these three variables, which leads to a missing value on the varibale I'm creating. I would like to find a way to make Stata calculate the mean with what is available, meaning:
-If there are no missing values, calulate the mean using all three observations
-If one observation is missing, calculate the mean of the two observations that are available
-If two observations are missing, make the observation on the new variabe equal to the observation from the variable that isn't missing

Is there any way I could code for that?

[Joro Kolev]

-egen, rowmean- does what you want.

[Carlo Lazzaro]

Aline:
as an aside to Joro's helpful advice, please note that Stata applies listwise deletion by default; hence, observations with missing value will be ruled out from subsequent calculation, as you can see from the following toy-example:

. use "C:\Program Files\Stata16\ado\base\a\auto.dta"
(1978 Automobile Data)

. sum rep78

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
       rep78 |         69    3.405797    .9899323          1          5

. sum price

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
       price |         74    6165.257    2949.496       3291      15906

.

[Aline Campos]

Wonderful, thank you both!