完全匹配某些列，部分匹配 inner_join

Question

I have two dataframes from different sources that refer to the same people, but due to errors from self-reported data, the dates may be slightly off.

Example data:

df1 <- data.frame(name= c("Ann", "Betsy", "Charlie", "Dave"), 
                  dob= c(as.Date("2000-01-01", "%Y-%m-%d"), as.Date("2001-01-01", "%Y-%m-%d"), 
                         as.Date("2002-01-01", "%Y-%m-%d"), as.Date("2003-01-01", "%Y-%m-%d")), 
                  stringsAsFactors=FALSE)

df2 <- data.frame(name= c("Ann", "Charlie", "Elmer", "Fred"), 
                  dob= c(as.Date("2000-01-11", "%Y-%m-%d"), as.Date("2004-01-01", "%Y-%m-%d"), 
                         as.Date("2001-01-01", "%Y-%m-%d"), as.Date("2006-01-01", "%Y-%m-%d")), 
                  stringsAsFactors=FALSE)

I want to match by exact name, with dplyr like:

library(dplyr)
inner_join(df1, df2, by = c("name"))

name    dob.x   dob.y
Ann     2000-01-01  2000-01-11
Charlie     2002-01-01  2004-01-01

but also by dates of birth (dob) within 30 days, with the fuzzyjoin package like:

library(fuzzyjoin)

difference_inner_join(df1, df2, by=c("dob"), max_dist = 30)

name.x  dob.x   name.y  dob.y
Ann     2000-01-01  Ann     2000-01-11
Betsy   2001-01-01  Elmer   2001-01-01

But combine both criteria, so that only Ann would be returned

Answer 1

Relying on dplyr and base R alone. I rarely need fuzzy joins. inner_join and subsequently filter usually is enough

inner_join(df1, df2, by = c("name")) %>% 
  filter(abs(difftime(dob.x,dob.y, units = "days"))<30)

result

  name      dob.x      dob.y
 1  Ann 2000-01-01 2000-01-11

Answer 2

Well you could do this:

 difference_inner_join(df1, df2, by=c("dob"), max_dist = 30) %>%
  filter(name.x == name.y)

  name.x      dob.x name.y      dob.y
1    Ann 2000-01-01    Ann 2000-01-11