Split data by focal column and remove duplicates from a span to columns, keeping the one with the most data

maxDup_NAs.R

# Create a dataframe of example data (here derived from raptor growth metrics)
df <- matrix(c(24,NA,365,1,NA,NA,6,33,NA,10,59,NA,37,300,477,NA,233,312,NA,NA,450,4,28,49),byrow = T, 8,3)
colnames(df) <- c("age","wing","mass")
df <- as.data.frame(df)

df$loc=c("Flev","Flev","Flev","Ters","Ters","Schi","Schi","Schi")
df$yr=c("2004","2004","2004","2007","2007","2004","2004","2008")
df

# split data by focal column and remove duplicates from a span to columns, keeping the one with the most data
res <- do.call(rbind,lapply(split(df, df$loc), 
                            function(x) x[which.max(rowSums(!is.na(x[1:3]))),]))
row.names(res) <- NULL
res

# Note that this is not perfect. In this example, there are two rows with two cells containing data for 
# location 'Flav'. R has chosen to retain the row with the greatest total value (re: `which.max`)