以长数据格式创建聚合变量

Question

I'm sure there's a question similar to this already, but I couldn't make them work

I am trying to calculate aggregates (or subtotals) in a dataframe of long format. In the group column I want an aggregate variable "AGG" that is a sum of "value" for a specific "Year" and "var". I have tried using the aggregate() function, but didn't succeed. I used the code:

aggregate(value ~ cbind(Year,var), data = Energi5, FUN = sum)

My data looks like this

> head(df)
     Year group  var     value
1    1966       A x   25465462
2    1966       B x    9512621
3    1966       E x    2832865
4    1966       H x     291769
5    1966      NE x  141524912
6    1966      NF x   23580353
> tail(df)
     Year group   var  value
5403 2017     NZ y    167158
5404 2017      O y     23480
5405 2017     QF y         0
5406 2017     QS y         0
5407 2017     QZ y     16447
5408 2017 TC3000 y    488556

and I would like to obtain something like this at the end of (or in the middle of) my existing dataframe

     Year group   var  value
5409 1966   AGG   x        ?
5410 1967   AGG   x        ?
...
5450 2017   AGG   x        ?
5451 1966   AGG   y        ?
...

I hope you can help. Thank you!

Answer 1

The error lies in how are you declaring the formula. See ?formula in the manual.

# Example
year <- rep(seq(1966, 2020), each = 8)
group <- rep(letters[1:4], times = 2*(2021-1966))
var <- rep(c("x", "y"), times = length(year)/2)
value <- rnorm(length(year))

data <- cbind.data.frame(year, group, var, value)

# Solution
aggregate(value ~ year * var, data, FUN=sum)

Answer 2

There is probably a more efficient way to do this, but does this help?

library(dplyr)

df <- Energi5 %>%  group_by(Year, var) %>% mutate(value = sum(value)) %>% summarise_all(funs(mean))

df$group <- "AGG"

Energi5 <- merge(Energi5, df, all = T)