calculate returns in data frame containing NA

In my case, NA means market was closed for holidays.
It doesn't occurs a lot (a few times /year).
I am OK to dismiss them, but is there any argument when using the diff(log)
function? I would like to avoid calculating a return between 2 sessions with
one day off between these 2 days, especially when this day off is not a
worldwide holyday.


J.E.T. SA
1 Place de Saint-Gervais
1201 Geneva, Switzerland
Tel: (41 22) 346 00 55
Fax: (41 22) 346 00 33
IM & BBM :arnaud.gaboury at gmail.com
E-mail: ag at jet-sa.ch

************************************************
- Secured E-mail with SSL protocol
- Encrypted IM with Simplite for Window Messenger