数据表总和和子集

Question

I have a data.table that I am wanting to aggregate

library(data.table)
dt1 <- data.table(year=c("2001","2001","2001","2002","2002","2002","2002"),
                  group=c("a","a","b","a","a","b","b"), 
                  amt=c(20,40,20,35,30,28,19))

I am wanting to sum the amt by year and group and then filter where the summed amt for any given group is greater than 100.

I've got the data.table sum nailed.

dt1[, sum(amt),by=list(year,group)]

   year group V1
1: 2001     a 60
2: 2001     b 20
3: 2002     a 65
4: 2002     b 47

I am having trouble with my final level of filtering.

The end outcome I am looking for is:

   year group V1
1: 2001     a 60
2: 2002     a 65

As a) 60 + 65 > 100 whereas b) 20 + 47 <= 100

Any thoughts on how to achieve this would be great.

I had a look at this data.table sum by group and return row with max value and was wondering whether or not their is an equally eloquent solution to my problem.

Answer 1

Single liner in data.table :

dt1[, lapply(.SD,sum), by=.(year,group)][, if (sum(amt) > 100) .SD, by=group]

#   group year amt
#1:     a 2001  60
#2:     a 2002  65

Answer 2

You can do:

library(dplyr)
dt1 %>% 
  group_by(group, year) %>% 
  summarise(amt = sum(amt)) %>%
  filter(sum(amt) > 100)

Which gives:

#Source: local data table [2 x 3]
#Groups: group
#
#  year group amt
#1 2001     a  60
#2 2002     a  65

Answer 3

This might not be an idea solution, but I would do that in several steps as follows:

dt2=dt1[, sum(amt),by=list(year,group)]
dt3=dt1[, sum(amt)>100,by=list(group)]
dt_result=dt2[group %in% dt3[V1==TRUE]$group,]

Answer 4

Here's a two-liner. Find the subset of groups you want first

big_groups <- dt1[,sum(amt),by=group][V1>100]$group
dt1[group%in%big_groups,sum(amt),by=list(year,group)]