R SummaryBy package grouping

Question

Have a csv file with the columns ABCDE

Created a fun1 function like this to summarize data

fun1 <- function(x){c(len=length(x), min=min(x), max=max(x))}

When I summarize on a particular column, it works

summaryBy(A ~ B, data=data1, FUN=fun1 , keep.names=TRUE)

But, How do I add a additional function in fun1 like sum(C) (Which is not relevant to x) and use it in summaryBy to get the relevant results for a groupBy of B?

For example,

A B C D E

1 2 3 4 5

1 2 4 5 7

1 3 5 7 8

Need to group by B (with respect to A), so will get two groups (2,3). But sum(c) will be irrespective of A.

Results should be

B len min max sum(c)

2  2   1   1   7

3  1   1   1   5

Answer 1

Try this:

summaryBy(A + C ~ B, data = data1, FUN = c(length, min, max, sum))[c(-3, -5, -7, -8)]

giving:

  B A.length A.min A.max C.sum
1 2        2     1     1     7
2 3        1     1     1     5

summaryBy might not be the best fit for that problem. With sqldf it could be written like this:

library(sqldf)
sqldf("select B, count(A) len, min(A) min, max(A) max, sum(C) sum from data1 group by B")

giving:

  B len min max sum
1 2   2   1   1   7
2 3   1   1   1   5

Note: In the examples above we used;

data1 <- structure(list(A = c(1L, 1L, 1L), B = c(2L, 2L, 3L), C = 3:5, 
    D = c(4L, 5L, 7L), E = c(5L, 7L, 8L)), .Names = c("A", "B", 
"C", "D", "E"), class = "data.frame", row.names = c(NA, -3L))