I have an example data frame below. I need to have it sorted by Type, Species, and BdFt. The example below is nearly correct, however I don't want species sorted in alphabetical order. I would like to sort species based on the sum of each species (within each "type") in descending order. So for example, within type 4404, 'DF' should appear first. I also would like to maintain the current number of observations, so I don't want to consolidate by species group. Could anyone help me achieve this, perhaps with dplyr?
-Brandon
Type Species BdFt
4404 BB 164
4404 BB 55
4404 BM 831
4404 BM 419
4404 BM 242
4404 BM 20
4404 CH 565
4404 CH 206
4404 CH 88
4404 CO 1817
4404 CO 531
4404 CO 286
4404 CO 31
4404 DF 19740
4404 DF 5930
4404 DF 613
4404 DF 468
4404 DF 167
4404 GF 360
4404 GF 232
4404 GF 124
4410 BM 909
4410 CH 161
4410 DF 18756
4410 GF 3642
4410 RA 549
Here is one option with arrange
library(dplyr)
df2 <- df1 %>%
arrange(Type, desc(ave(BdFt, Species, Type, FUN = sum)))
df2
# Type Species BdFt
#1 4404 DF 19740
#2 4404 DF 5930
#3 4404 DF 613
#4 4404 DF 468
#5 4404 DF 167
#6 4404 CO 1817
#7 4404 CO 531
#8 4404 CO 286
#9 4404 CO 31
#10 4404 BM 831
#11 4404 BM 419
#12 4404 BM 242
#13 4404 BM 20
#14 4404 CH 565
#15 4404 CH 206
#16 4404 CH 88
#17 4404 GF 360
#18 4404 GF 232
#19 4404 GF 124
#20 4404 BB 164
#21 4404 BB 55
#22 4410 DF 18756
#23 4410 GF 3642
#24 4410 BM 909
#25 4410 RA 549
#26 4410 CH 161
Or with order
from base R
df1[with(df1, order(Type, -ave(BdFt, Species, Type, FUN = sum))),]
df1 <- structure(list(Type = c(4404L, 4404L, 4404L, 4404L, 4404L, 4404L,
4404L, 4404L, 4404L, 4404L, 4404L, 4404L, 4404L, 4404L, 4404L,
4404L, 4404L, 4404L, 4404L, 4404L, 4404L, 4410L, 4410L, 4410L,
4410L, 4410L), Species = c("BB", "BB", "BM", "BM", "BM", "BM",
"CH", "CH", "CH", "CO", "CO", "CO", "CO", "DF", "DF", "DF", "DF",
"DF", "GF", "GF", "GF", "BM", "CH", "DF", "GF", "RA"), BdFt = c(164L,
55L, 831L, 419L, 242L, 20L, 565L, 206L, 88L, 1817L, 531L, 286L,
31L, 19740L, 5930L, 613L, 468L, 167L, 360L, 232L, 124L, 909L,
161L, 18756L, 3642L, 549L)), class = "data.frame", row.names = c(NA,
-26L))