CODEWITHSUNDEEP
rdataframedplyrtidyverse
coding
coding

Reputation: 1247

Group by two column and summarize multiple columns

I have a data frame and I would like to group by the column "State" and "Date" and then summarize the values of the other columns something like this.

df

State  Female  Male   Date
------------------------------
Texas  2       2     01/01/04
Texas  3        1     01/01/04
Texas  5        4     02/01/04
Cali   1        1     05/06/05
Cali   2        1     05/06/05
Cali   3         1    10/06/05
Cali   1         2     10/06/05
NY    10         5    11/06/05
NY    11         6    12/06/05

Outcome expected

df

State  Female  Male   Date
------------------------------
Texas  5       3     01/01/04
Texas  5        4     02/01/04
Cali   3        2     05/06/05
Cali   4         3    10/06/05
NY    10         5    11/06/05
NY    11         6    12/06/05

I tried with group by and summarize but I don´t exactly how con I do the same for 2 columns

My try

df <- df_homicides %>% 
        group_by(state) %>% 
        summarise(Female = sum(Female))

``
Thanks for your help!

Upvotes: 4

Views: 9341

Answers (2)

akrun
akrun

Reputation: 887981

We can use summarise with across from dplyr version > = 1.00

library(dplyr)
df %>%
   group_by(State, Date) %>%
   summarise(across(everything(), sum, na.rm = TRUE), .groups = 'drop')
# A tibble: 6 x 4
#  State Date       Female  Male
#  <chr> <chr>       <int> <int>
#1 Cali  05/06/2005      3     2
#2 Cali  10/06/2005      4     3
#3 NY    11/06/2005     10     5
#4 NY    12/06/2005     11     6
#5 Texas 01/01/2004      5     3
#6 Texas 02/01/2004      5     4

Or using aggregate from base R

aggregate(.~ State + Date, df, sum, na.rm = TRUE)

data

df <-  structure(list(State = c("Texas", "Texas", "Texas", "Cali", "Cali", 
"Cali", "Cali", "NY", "NY"), Female = c(2L, 3L, 5L, 1L, 2L, 3L, 
1L, 10L, 11L), Male = c(2L, 1L, 4L, 1L, 1L, 1L, 2L, 5L, 6L), 
    Date = c("01/01/2004", "01/01/2004", "02/01/2004", "05/06/2005", 
    "05/06/2005", "10/06/2005", "10/06/2005", "11/06/2005", "12/06/2005"
    )), class = "data.frame", row.names = c(NA, -9L))

Upvotes: 3

Duck
Duck

Reputation: 39623

Try this. You can use summarise_all() to aggregate multiple variables with a desired function. Here the code:

library(dplyr)
#Code
df %>% group_by(State,Date) %>%
  summarise_all(.funs = sum,na.rm=T)

Output:

# A tibble: 6 x 4
# Groups:   State [3]
  State Date       Female  Male
  <chr> <chr>       <int> <int>
1 Cali  05/06/2005      3     2
2 Cali  10/06/2005      4     3
3 NY    11/06/2005     10     5
4 NY    12/06/2005     11     6
5 Texas 01/01/2004      5     3
6 Texas 02/01/2004      5     4

Some data used:

#Data
df <- structure(list(State = c("Texas", "Texas", "Texas", "Cali", "Cali", 
"Cali", "Cali", "NY", "NY"), Female = c(2L, 3L, 5L, 1L, 2L, 3L, 
1L, 10L, 11L), Male = c(2L, 1L, 4L, 1L, 1L, 1L, 2L, 5L, 6L), 
    Date = c("01/01/2004", "01/01/2004", "02/01/2004", "05/06/2005", 
    "05/06/2005", "10/06/2005", "10/06/2005", "11/06/2005", "12/06/2005"
    )), class = "data.frame", row.names = c(NA, -9L))

Upvotes: 1

Related Questions