[R] Counting the number of non-NA values per day

Tim Chatterton tim.chatterton at uwe.ac.uk
Tue Aug 11 19:26:14 CEST 2009

I have a long dataframe ("pollution") that contains a column of hourly 
date information ("date") and  a column of  pollution  measurements ("pol")

I have been happily calculating daily means and daily maximums using the 
aggregate function

 DMEANpollution =  aggregate(pollution["pol"], 
format(pollution["date"],"%Y-%j"), mean, na.rm = TRUE)
 DMAXpollution =  aggregate(pollution["pol"], 
format(pollution["date"],"%Y-%j"), max, na.rm = TRUE)

However, I also need to count the number of valid measurements for each 
day to check that the mean and max are likely to be valid (for example I 
need at least 18 hourly measurements to calculate a valid daily mean)

Try as I might I have not found a simple way of doing this.
Can anybody help please?

Many thanks,



Dr Tim Chatterton
Senior Research Fellow
Air Quality Management Resource Centre
Faculty of Environment and Technology
University of the West of England
Frenchay Campus
BS16 1QY

Tel: 0117 328 2929
Fax: 0117 328 3360
Email: Tim.Chatterton at uwe.ac.uk

More information about the R-help mailing list