R 在尽可能多的列上串行连接两个数据帧，直到df1中的每一行都匹配为止

我有两个数据帧需要连接。目标是连接尽可能多的列，并从最多到最少的精确列匹配顺序进行。我需要解决K列的问题。对于匹配j列（j

# 3-column toy example 
library(dplyr)
df1_rows = 50
df2_rows = 100
set.seed(1)

df1 <- data.frame(id1 = sample(LETTERS,df1_rows,replace = T),
                  A = sample(1:5,df1_rows,replace = T),
                  B = sample(1:5,df1_rows,replace = T),
                  C = sample(1:5,df1_rows,replace = T) )
df1
df2 <- data.frame(A = sample(1:5,df2_rows,replace = T),
                  B = sample(1:5,df2_rows,replace = T),
                  C = sample(1:5,df2_rows,replace = T) ) %>%
  distinct(A,B,C,.keep_all = T) %>%
  distinct(A,B,.keep_all = T) %>%
  distinct(B,C,.keep_all = T) %>%
  distinct(A,C,.keep_all = T) %>%
  rowwise() %>%
  mutate(id2 = sample(LETTERS,1,replace = T))

# match on 3 columns
col_match_3 = df1 %>%
  inner_join(df2 , by = c("A","B","C")) %>%
  distinct(id1, A,B,C,.keep_all = T)

#serial matches on two columns
col_match_AB = df1 %>%
  anti_join(col_match_3, by = c("A","B","C")) %>%
  inner_join(df2 %>% select(-C), by = c("A","B")) %>%
  distinct(id1, A,B,C,.keep_all = T)

col_match_BC = df1 %>%
  anti_join(col_match_3, by = c("A","B","C")) %>%
  anti_join(col_match_AB, by = c("A","B")) %>%
  inner_join(df2 %>% select(-A), by = c("B","C")) %>%
  distinct(id1, A,B,C,.keep_all = T)

col_match_AC = df1 %>%
  anti_join(col_match_3, by = c("A","B","C")) %>%
  anti_join(col_match_AB, by = c("A","B")) %>%
  anti_join(col_match_BC, by = c("B","C")) %>%
  inner_join(df2 %>% select(-B), by = c("A","C")) %>%
  distinct(id1, A,B,C,.keep_all = T)

#join on one col
col_match_A = df1 %>%
  anti_join(col_match_3, by = c("A","B","C")) %>%
  anti_join(col_match_AB, by = c("A","B")) %>%
  anti_join(col_match_BC, by = c("B","C")) %>%
  anti_join(col_match_AC, by = c("A","C")) %>%
  inner_join(df2 %>% select(id2,A), by = "A") %>%
  distinct(id1, A,B,C,.keep_all = T)

col_match_B = df1 %>%
  anti_join(col_match_3, by = c("A","B","C")) %>%
  anti_join(col_match_AB, by = c("A","B")) %>%
  anti_join(col_match_BC, by = c("B","C")) %>%
  anti_join(col_match_AC, by = c("A","C")) %>%
  anti_join(col_match_A, by = "A") %>%
  inner_join(df2 %>% select(id2,B), by = "B") %>%
  distinct(.keep_all = T)
  

res = 
  col_match_3 %>%
  bind_rows(col_match_AB) %>%
  bind_rows(col_match_BC) %>%
  bind_rows(col_match_AC) %>%
  bind_rows(col_match_A)

res

#3列玩具示例
图书馆（dplyr）
df1_行=50
df2_行=100
种子（1）
df1%
不同的（A，B，.keep_all=T）%>%
不同的（B，C，.keep_all=T）%>%
不同的（A，C，.keep_all=T）%>%
行（）
变异（id2=样本（字母，1，替换=T））
#在3列上匹配
列匹配3=df1%>%
内部联接（df2，by=c（“A”、“B”、“c”））%>%
不同（id1、A、B、C、.keep_all=T）
#两列上的序列匹配
列匹配=df1%>%
反联接（列匹配3，由=c（“A”、“B”、“c”））%>%
内部联接（df2%>%select（-C），by=C（“A”，“B”））%>%
不同（id1、A、B、C、.keep_all=T）
col_match_BC=df1%>%
反联接（列匹配3，由=c（“A”、“B”、“c”））%>%
反连接（col\u match\u AB，by=c（“A”，“B”））%>%
内部联接（df2%>%select（-A），by=c（“B”，“c”））%>%
不同（id1、A、B、C、.keep_all=T）
列匹配=df1%>%
反联接（列匹配3，由=c（“A”、“B”、“c”））%>%
反连接（col\u match\u AB，by=c（“A”，“B”））%>%
反连接（col\u match\u BC，by=c（“B”，“c”））%>%
内部联接（df2%>%select（-B），by=c（“A”，“c”））%>%
不同（id1、A、B、C、.keep_all=T）
#连成一列
列匹配=df1%>%
反联接（列匹配3，由=c（“A”、“B”、“c”））%>%
反连接（col\u match\u AB，by=c（“A”，“B”））%>%
反连接（col\u match\u BC，by=c（“B”，“c”））%>%
反连接（col\u match\u AC，by=c（“A”，“c”））%>%
内部联接（df2%%>%select（id2，A），by=“A”）%%>%
不同（id1、A、B、C、.keep_all=T）
列匹配=df1%>%
反联接（列匹配3，由=c（“A”、“B”、“c”））%>%
反连接（col\u match\u AB，by=c（“A”，“B”））%>%
反连接（col\u match\u BC，by=c（“B”，“c”））%>%
反连接（col\u match\u AC，by=c（“A”，“c”））%>%
反联接（列匹配A，by=“A”）%>%
内部联接（df2%%>%select（id2，B），by=“B”）%%>%
不同（.keep_all=T）
res=
列匹配\u 3%>%
绑定行（列匹配对象）%>%
绑定行（列匹配）%>%
绑定行（列匹配）%>%
绑定行（列匹配）
物件