将 Dataframe 转换为 R 中的二进制矩阵

Question

I have a data frame with 2 columns, customerID and StockCodes describing stockcodes bought by the customer over a period of time. There could be multiple observations for the same customer as he might have bought the same items multiple times or different items over a period of time. The sample data looks as follows:

CustomerID StockCode

12346 23166

12347 16008

12347 17021

12347 20665

12347 20719

12347 20719

12347 20719

12347 20719

12347 20780

12347 20782

12347 20966

12347 21035

I need to transpose the data frame in R such that all stockcodes would appear as columns without any repetition and each row will have a distinct customerID. I have two questions:

1. The cross-section cell value will have either numeric '1' if the customer has at least one matching stock code else 0.

2. The cross-section cell will have the count of stockcodes each customer has, if there is a matching stock code, else 0.

Answer 1

This is easily done with dplyr and tidyr::pivot_wider .

Data

example <- data.frame(CustomerID = c(12346, 12347, 12347, 12347, 12347, 12347), 
                      StockCode = c(23166, 16008, 17021, 20665, 20719, 20719)
)

Code for Part (1)

 A <- example %>% 
    distinct %>%
      mutate(Test = 1) %>% 
        tidyr::pivot_wider(values_from = Test, names_from = StockCode) %>% 
            replace(is.na(.), 0)

Output for Part (1)

# A tibble: 2 x 6
  CustomerID `23166` `16008` `17021`
       <dbl>   <dbl>   <dbl>   <dbl>
1      12346       1       0       0
2      12347       0       1       1
# ... with 2 more variables:
#   `20665` <dbl>, `20719` <dbl>

Code for Part (2)

B <- example %>% 
  group_by_all %>%
    count %>% 
     tidyr::pivot_wider(values_from = n, names_from = StockCode) %>% 
       replace(is.na(.), 0)

Output for Part (2)

> B
# A tibble: 2 x 6
# Groups:   CustomerID [2]
  CustomerID `23166` `16008` `17021`
       <dbl>   <int>   <int>   <int>
1      12346       1       0       0
2      12347       0       1       1
# ... with 2 more variables:
#   `20665` <int>, `20719` <int>