st: Re: egen rowmin rowmax

[Kit Baum <baum@bc.edu>]

This should not be necessary.


        rowtotal(varlist)
            may not be combined with by.  It creates the (row) sum  
of the
            variables in varlist, treating missing as 0.

        rowmin(varlist)
            may not be combined with by.  It gives the minimum value  
in varlist
            for each observation (row).  If all values in varlist  
are missing for
            an observation, newvar is set to missing.


The egen rowtotal function will properly handle missings, and it  
properly handles varlists, including
the form firstvar-lastvar. (Remember that this is not the alpha order  
of your vars, but the data set order) Thus

webuse auto
egen tot = rowtotal(price-headroom)

will produce a variable with 74 obs. even though rep78 is in there.   
You refer to an egen function total(), which  produces something  
quite different: a constant which is the sum of an exp. An exp of,  
say, price-headroom is exactly that: subtract headroom from price and  
sum it up. If MVs make the exp missing, then the sum will be missing.  
Perfectly logical: you get what you ask for. But you probably meant  
to use rowtotal.

Furthermore the rowmin function will properly pick out the minimum  
value, ignoring missings, unless all are
missing, so it should return zero for a row with all (or some)  
zeroes, greater than zero otherwise, if none of your rows contain all  
MVs.

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


On Nov 5, 2006, at 2:33 AM, statalist-digest wrote:


> As it turns out, egen [var] = rowsum([varlist]) is equivalent to
> egen [var] = total([varlist]).
>
> My solution turned out to be to replace all of the zeros  in the seq*
> cells with missings, then execute
> egen byte minmath = rowmin(seq*) /*without the condition!*/
> and add the following line of code:
> replace minmath = 0 if minmath==.
*
*   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/