基于一列的完全匹配和两列的模糊匹配对两个数据帧的内部连接

Question

我想对我的一列 (Product_date) 执行精确匹配，然后对 product_name 和 state_name 进行部分匹配或模糊匹配。

例如：

df1 <- data.frame(ID=c("P01", "P04", "P23"),
                  Product_name=c("Jewel", "Bronze", "Iron"), 
                  Product_state=c("Kansas", "Illinois", "Florida"),
                  Product_date=c("2021-08-01", "2021-01-01", "2020-12-21"))

df2 <- data.frame(
  Product_name=c("Jewel", "Bro", "Ir", "Uknw"), 
  Product_state=c("Kansasss", "IllI", "Flor_ida", "Cali2"),
  Product_date=c("2021-08-01", "2021-01-01", "2020-12-21", "2020-09"),
  Product_status=c("sold", "lost", "sold", "sold"))

desired_df <-  data.frame(c("P01", "P04", "P23"),
                          Product_name=c("Jewel", "Bronze", "Iron"), 
                          Product_state=c("Kansas", "Illinois", "Florida"),
                          Product_date=c("2021-08-01", "2021-01-01", "2020-12-21"), 
                          Product_name=c("Je", "Bro", "Ir"), 
                          Product_state=c("Kansasss", "IllI", "Flor_ida"),
                          Product_date=c("2021-08-01", "2021-01-01", "2020-12-21"), 
                          Product_status=c("sold", "lost", "sold"))

仅出于说明目的，这就是我脑海中的代码的样子（但当然它不起作用）

matched <- df1 %>%
stringdist_inner_join(df2, by= c("Product_name", max_dist=2),
                           by= c("Product_stat", max_dist=4), 
                           by = c("Product_date"))

Answer 1

一个可能的解决方案：

library(fuzzyjoin)
library(dplyr)

stringdist_join(df1, df2, 
                by = c("Product_name","Product_state"),
                mode = "left",
                ignore_case = FALSE, 
                method = "jw", 
                max_dist = 0.5) %>% 
  filter(Product_date.x == Product_date.y)
#>    ID Product_name.x Product_state.x Product_date.x Product_name.y
#> 1 P01          Jewel          Kansas     2021-08-01          Jewel
#> 2 P04         Bronze        Illinois     2021-01-01            Bro
#> 3 P23           Iron         Florida     2020-12-21             Ir
#>   Product_state.y Product_date.y Product_status
#> 1        Kansasss     2021-08-01           sold
#> 2            IllI     2021-01-01           lost
#> 3        Flor_ida     2020-12-21           sold