aggregate_by aggregates some variables of a data frame according to
the values of a categorical variable of the same data frame and using a given
function for the aggregation.
aggregate_by(df, col_name, ..., .funs = sum)
| df | the data frame on which to perform the aggregation. |
|---|---|
| col_name | the variable on which to perform the aggregation |
| ... | other variables on which to perform the aggregation |
| .funs | the function used to perform the aggregation.
By default |
A data frame with the same variables as `df` but for which some of the observation have been aggregated (i.e. less rows than in `df`).
On a data frame that contains the values of the categorical variable to
aggregate (`col_name`, `...`), the function performs a
group_by followed by a summarise
with the function(s) .funs.
library(dplyr) ## A toy data frame: set.seed(30101976) data <- expand.grid(letters[1:3], 1:3, 4:6) data$Var4 <- sample(1:100, nrow(data), TRUE) data$Var5 <- sample(1:100, nrow(data), TRUE) data$Var6 <- sample(1:100, nrow(data), TRUE) ## Aggregating the values "a" and "b" of the categorical variable Var1, ## summing the values of variables Var4, Var5, Var6 (i.e. all the variables ## that are not in the arguments of the function call): data %>% mutate(Var1 = recode(Var1, a = "b")) %>% aggregate_by(Var1, Var2, Var3)#> [90m# A tibble: 18 x 6[39m #> [90m# Groups: Var1, Var2 [?][39m #> Var1 Var2 Var3 Var4 Var5 Var6 #> [3m[90m<fct>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m #> [90m 1[39m b 1 4 77 136 142 #> [90m 2[39m b 1 5 119 116 106 #> [90m 3[39m b 1 6 14 111 107 #> [90m 4[39m b 2 4 77 122 117 #> [90m 5[39m b 2 5 78 33 98 #> [90m 6[39m b 2 6 113 144 171 #> [90m 7[39m b 3 4 54 81 82 #> [90m 8[39m b 3 5 139 138 123 #> [90m 9[39m b 3 6 49 47 58 #> [90m10[39m c 1 4 49 12 62 #> [90m11[39m c 1 5 5 27 17 #> [90m12[39m c 1 6 41 10 43 #> [90m13[39m c 2 4 42 9 6 #> [90m14[39m c 2 5 27 35 52 #> [90m15[39m c 2 6 35 38 3 #> [90m16[39m c 3 4 16 2 96 #> [90m17[39m c 3 5 21 5 10 #> [90m18[39m c 3 6 28 27 7## To calculate the mean value: data %>% mutate(Var1 = recode(Var1, a = "b")) %>% aggregate_by(Var1, Var2, Var3, .funs = mean)#> [90m# A tibble: 18 x 6[39m #> [90m# Groups: Var1, Var2 [?][39m #> Var1 Var2 Var3 Var4 Var5 Var6 #> [3m[90m<fct>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<dbl>[39m[23m [3m[90m<dbl>[39m[23m [3m[90m<dbl>[39m[23m #> [90m 1[39m b 1 4 38.5 68 71 #> [90m 2[39m b 1 5 59.5 58 53 #> [90m 3[39m b 1 6 7 55.5 53.5 #> [90m 4[39m b 2 4 38.5 61 58.5 #> [90m 5[39m b 2 5 39 16.5 49 #> [90m 6[39m b 2 6 56.5 72 85.5 #> [90m 7[39m b 3 4 27 40.5 41 #> [90m 8[39m b 3 5 69.5 69 61.5 #> [90m 9[39m b 3 6 24.5 23.5 29 #> [90m10[39m c 1 4 49 12 62 #> [90m11[39m c 1 5 5 27 17 #> [90m12[39m c 1 6 41 10 43 #> [90m13[39m c 2 4 42 9 6 #> [90m14[39m c 2 5 27 35 52 #> [90m15[39m c 2 6 35 38 3 #> [90m16[39m c 3 4 16 2 96 #> [90m17[39m c 3 5 21 5 10 #> [90m18[39m c 3 6 28 27 7## or to apply it to all the columns: data %>% mutate(Var1 = recode(Var1, a = "b")) %>% aggregate_by(Var1, Var2, Var3, .funs = list(sum, mean))#> [90m# A tibble: 18 x 9[39m #> [90m# Groups: Var1, Var2 [?][39m #> Var1 Var2 Var3 Var4_sum Var5_sum Var6_sum Var4_mean Var5_mean Var6_mean #> [3m[90m<fct>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<dbl>[39m[23m [3m[90m<dbl>[39m[23m [3m[90m<dbl>[39m[23m #> [90m 1[39m b 1 4 77 136 142 38.5 68 71 #> [90m 2[39m b 1 5 119 116 106 59.5 58 53 #> [90m 3[39m b 1 6 14 111 107 7 55.5 53.5 #> [90m 4[39m b 2 4 77 122 117 38.5 61 58.5 #> [90m 5[39m b 2 5 78 33 98 39 16.5 49 #> [90m 6[39m b 2 6 113 144 171 56.5 72 85.5 #> [90m 7[39m b 3 4 54 81 82 27 40.5 41 #> [90m 8[39m b 3 5 139 138 123 69.5 69 61.5 #> [90m 9[39m b 3 6 49 47 58 24.5 23.5 29 #> [90m10[39m c 1 4 49 12 62 49 12 62 #> [90m11[39m c 1 5 5 27 17 5 27 17 #> [90m12[39m c 1 6 41 10 43 41 10 43 #> [90m13[39m c 2 4 42 9 6 42 9 6 #> [90m14[39m c 2 5 27 35 52 27 35 52 #> [90m15[39m c 2 6 35 38 3 35 38 3 #> [90m16[39m c 3 4 16 2 96 16 2 96 #> [90m17[39m c 3 5 21 5 10 21 5 10 #> [90m18[39m c 3 6 28 27 7 28 27 7## To calculate the mean and the sum: data %>% mutate(Var1 = recode(Var1, a = "b")) %>% aggregate_by(Var1, Var2, Var3, .funs = list(sum(Var6), mean(Var4)))#> [90m# A tibble: 18 x 5[39m #> [90m# Groups: Var1, Var2 [?][39m #> Var1 Var2 Var3 `sum(Var6)` `mean(Var4)` #> [3m[90m<fct>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<int>[39m[23m [3m[90m<dbl>[39m[23m #> [90m 1[39m b 1 4 142 38.5 #> [90m 2[39m b 1 5 106 59.5 #> [90m 3[39m b 1 6 107 7 #> [90m 4[39m b 2 4 117 38.5 #> [90m 5[39m b 2 5 98 39 #> [90m 6[39m b 2 6 171 56.5 #> [90m 7[39m b 3 4 82 27 #> [90m 8[39m b 3 5 123 69.5 #> [90m 9[39m b 3 6 58 24.5 #> [90m10[39m c 1 4 62 49 #> [90m11[39m c 1 5 17 5 #> [90m12[39m c 1 6 43 41 #> [90m13[39m c 2 4 6 42 #> [90m14[39m c 2 5 52 27 #> [90m15[39m c 2 6 3 35 #> [90m16[39m c 3 4 96 16 #> [90m17[39m c 3 5 10 21 #> [90m18[39m c 3 6 7 28