r - Creating a new column where each element is the count of subsets of two other columns, loop free -

this question has answer here:

• count number of observations/rows per group , add result data frame 7 answers

i have data frame looks table below keeps track of person visiting store in month. want create new column, total_visits, count of number of times id visited store during month. in below example, date 6-13 , id 23, total_visits have 3 in row date == 6-13, , id == 23.

date    id    6-13    23    6-13    34    6-13    23    6-13    23    7-13    23    

data frame i'm looking

date    id    total_visits    6-13    23    3 6-13    34    1 6-13    23    3 6-13    23    3 7-13    23    1 

while assume there sort of acast function ensure don't have loop through (30,000 rows), ok loop if vectorization did not work.

you can use dplyr package:

library(dplyr) df %>%  group_by(date, id) %>% mutate(total_visits = n())  # # tibble: 5 x 3  # # groups:   date, id [3]  #     date    id total_visits  #   <fctr> <int>        <int>  # 1   6-13    23            3  # 2   6-13    34            1  # 3   6-13    23            3  # 4   6-13    23            3  # 5   7-13    23            1 

use data.frame on output make dataframe.

update:

or using data.table package:

library(data.table) setdt(df)[, total_visits:=.n, by=c("date","id")]  df  #    date id total_visits  # 1: 6-13 23            3  # 2: 6-13 34            1  # 3: 6-13 23            3  # 4: 6-13 23            3  # 5: 7-13 23            1 

data:

df <- structure(list(date = structure(c(1l, 1l, 1l, 1l, 2l), .label = c("6-13",        "7-13"), class = "factor"), id = c(23l, 34l, 23l, 23l, 23l)), .names = c("date",        "id"), class = "data.frame", row.names = c(na, -5l))