I've used the arrange and mutate combination to do additions based on groupings. For example, I've used the following:
master_df <-group_by(master_df,asof_dt)
mutate(master_df,tot_flag=ls_flag)
This groups my data frame master_df by asof_dt, and then creates tot_flag and adds ls_flag by date.
However, my ls_flag column contains NA's.
I would like to do the following: 1) find out how to add the ls_flag, ignoring any NA's 2) find out how to add the total number of NA's per day.
Here is the full example:
asof_dt<-c("2014-10-01","2014-10-01","2014-10-01","2014-10-02","2014-10-02","2014-10-02")
ls_flag<-c(1,1,NA,NA,1,1)
master_df<-data.frame(asof_dt,ls_flag)
master_df <-group_by(master_df,asof_dt)
mutate(master_df,tot_flag=sum(ls_flag))
Thank you very much!
Is this your desired result? You can use the na.rm = TRUE option in sum():
master_df %>%
group_by(asof_dt) %>%
mutate(tot_flag = sum(ls_flag, na.rm = TRUE),
tot_NA = sum(is.na(ls_flag)))
#Source: local data frame [6 x 4]
#Groups: asof_dt
#
# asof_dt ls_flag tot_flag tot_NA
#1 2014-10-01 1 2 1
#2 2014-10-01 1 2 1
#3 2014-10-01 NA 2 1
#4 2014-10-02 NA 2 1
#5 2014-10-02 1 2 1
#6 2014-10-02 1 2 1
Or perhaps you just want a "summary" (using summarise):
master_df %>%
group_by(asof_dt) %>%
summarise(tot_flag = sum(ls_flag, na.rm = TRUE),
tot_NA = sum(is.na(ls_flag)))
#Source: local data frame [2 x 3]
#
# asof_dt tot_flag tot_NA
#1 2014-10-01 2 1
#2 2014-10-02 2 1