Impute missing observations R -

i have dataframe of calendar days , hourly observations associated values so:

   starttime  hour  delay    04-22      0     10    04-22      1     10    04-22      3     10    04-22      4     10 

what's needed hours of day , 0 associated said missing hours instead of no value or na. how best achieved? have attempted full_join dplyr dummy data frame, method seems clunky , inefficient.

in short, need this:

   starttime  hour  delay    04-22      0     10    04-22      1     10    04-22      2     0    04-22      3     10    04-22      4     10 

you use data.table efficiently join dataset. convert 'data.frame' 'data.table(setdt(df1)), set key columns (setkey(.., starttime, hour)`), join combination of unique values of 'starttime' , 'hour', , replace na values in delay '0'

library(data.table) setkey(setdt(df1), starttime, hour)[cj(starttime=unique(starttime),       hour=min(hour):max(hour))][is.na(delay), delay:=0l] #   starttime hour delay #1:     04-22    0    10 #2:     04-22    1    10 #3:     04-22    2     0 #4:     04-22    3    10 #5:     04-22    4    10 

or using merge/expand.grid base r, can above result

merge(expand.grid(starttime=unique(df1$starttime),       hour= min(df1$hour):max(df1$hour)), df1, all.x=true)