我有2个数据帧,并且我正在尝试使用条件进行内部联接。我将显示一个示例来阐明我要执行的操作:
A
cnumero cep numero bairro
1 124,35 124 35 K
2 375,NA 375 NA L
3 NA,28 NA 28 M
B
cnumero bairro XY
1 124,35 J 1
2 375,48 L 2
3 135,28 M 3
要合并的第一个条件是,如果A$cep != NA
和A$numero != NA
,则使用列cnumero
进行连接,否则使用列bairro
进行连接,结果:
new_A
cnumero cep numero XY
1 124,35 124 35 1
2 375,NA 375 NA 2
3 NA,28 NA 28 3
到目前为止,我所做的是使用此方法进行内部联接:A[A$cnumero %in% unique(B$cnumero),]
,因为在我的实际数据帧中,我在数据帧B
中有重复的值。
编辑:我的数据示例
A = data.frame(cnumero=c("124,35", "375,NA", "NA,28"),cep = c(124, 375, NA), numero = c(35, NA, 28), bairro = c("K", "L","M"))
B = data.frame(cnumero=c("124,35", "375,48", "135,28"), bairro = c("J", "L","M"), XY = c(1, 2, 3))
new_A = data.frame(cnumero=c("124,35", "375,NA", "NA,28"),cep = c(124, 375, NA), numero = c(35, NA, 28), XY = c(1, 2, 3))
以这种方式在基数R中的解决方案呢,分两个步骤,首先是第一个条件的连接,然后是第二个,最后是结果的汇总:
# the join with the first condition
A_1 <- merge(A[!is.na(A$cep) & !grepl('NA',A$cnumero), ],B, by = 'cnumero')
# select the column you need
A_1 <- A_1[,c("cnumero", "cep","numero","XY")]
# join for the second condition
A_2 <- merge(A[is.na(A$cep) | grepl('NA',A$cnumero), ],B, by = 'bairro')
# select columns you need
A_2 <- A_2[,c("cnumero.x", "cep","numero","XY")]
# rename the second part's columns
colnames(A_2) <- colnames(A_1)
# now the result
new_A <- rbind(A_1,A_2)
new_A
cnumero cep numero XY
1 124,35 124 35 1
2 375,NA 375 NA 2
3 NA,28 NA 28 3
# in case you need to remove the "temporary" tables
# remove(A_1, A_2)