CODEWITHSUNDEEP
rdataframedplyrpurrr
pontilicious
pontilicious

Reputation: 299

count rows per condition from another data frame

i am trying my hand at a private project with R. Following problem:

I have two data frames. Exemplary the tables of the two frames:

Frame1

          Home           Away
1         Lens       Paris SG
2 Rapid Vienna         Admira
3         LASK Austria Vienna
4 Shijiazhuang     Wuhan Zall
5  Sonderjyske    Midtjylland
6    Bohemians      Waterford

Frame2

# A tibble: 6 x 9
  Country League  Date       Home               Away                HG    AG Res      TG
  <chr>   <chr>   <chr>      <chr>              <chr>            <dbl> <dbl> <chr> <dbl>
1 Mexico  Liga MX 10/09/2020 Santos Laguna      U.N.A.M.- Pumas      1     2 A         3
2 Mexico  Liga MX 10/09/2020 Mazatlan FC        Club Tijuana         1     0 H         1
3 Mexico  Liga MX 10/09/2020 Cruz Azul          Pachuca              1     0 H         1
4 Mexico  Liga MX 09/09/2020 Club Leon          U.A.N.L.- Tigres     1     1 D         2
5 Mexico  Liga MX 09/09/2020 Puebla             Club America         2     3 A         5
6 Mexico  Liga MX 09/09/2020 Guadalajara Chivas Queretaro            1     1 D         2

now I want to insert a new column in the first data frame which excludes and counts the number of direct encounters from the second data frame, i.e. Home == Home Team and Away == AwayTeam. Is it possible to insert data in a dataframe that is linked to data from another dataframe?

Upvotes: 2

Views: 534

Answers (2)

akrun
akrun

Reputation: 886938

We can use data.table to do a join on the columns while getting the frequency count

library(data.table)
setDT(df_2)[df_1, .N, on = .(home, away), by = .EACHI]
#   home away N
#1:    a    c 2
#2:    b    a 0
#3:    c    b 1

Or using base R with table

df_1$Count <-  with(df_2, table(factor(paste(home, away),
        levels = unique(paste(df_1$home, df_1$away)))))

data

df_2 <- structure(list(home = c("a", "a", "c", "b"), away = c("c", "c", 
"b", "c")), class = "data.frame", row.names = c(NA, -4L))

df_1 <- structure(list(home = c("a", "b", "c"), away = c("c", "a", "b"
)), class = "data.frame", row.names = c(NA, -3L))

Upvotes: 2

starja
starja

Reputation: 10365

Yes, you can calculate the encounters in the second data.frame and then join this with the first data.frame:

df_1 <- data.frame(
  home = c("a", "b", "c"),
  away = c("c", "a", "b")
)

df_1
#>   home away
#> 1    a    c
#> 2    b    a
#> 3    c    b

df_2 <- data.frame(
  home = c("a", "a", "c", "b"),
  away = c("c", "c", "b", "c")
)

df_2
#>   home away
#> 1    a    c
#> 2    a    c
#> 3    c    b
#> 4    b    c

library(dplyr)

df_2_stats <- df_2 %>% 
  group_by(home, away) %>% 
  summarise(number_encounters = n())
#> `summarise()` regrouping output by 'home' (override with `.groups` argument)

df_2_stats
#> # A tibble: 3 x 3
#> # Groups:   home [3]
#>   home  away  number_encounters
#>   <chr> <chr>             <int>
#> 1 a     c                     2
#> 2 b     c                     1
#> 3 c     b                     1

df_1 <- df_1 %>% 
  left_join(df_2_stats, by = c("home", "away"))

df_1
#>   home away number_encounters
#> 1    a    c                 2
#> 2    b    a                NA
#> 3    c    b                 1

Created on 2020-09-11 by the reprex package (v0.3.0)

Upvotes: 0

Related Questions