R 当A列值匹配时，用B列值替换B列NAs

我试图基于两个变量（Entrez.ID和Gene.ID）合并两个数据帧。一个数据帧只有那些变量，例如

Entrez.ID  Gene.ID
10007      GNPDA1
10016      ALG2
10044      SH2D3C 

以及一个具有三个变量的数据帧，例如

Entrez.ID    Gene.ID   Ensembl.ID
10007        GPI       ENSG00000113552
10016        PDCD6     ENSG00000249915
10044        CHAT      ENSG00000095370

当前，当我使用以下命令合并文件时：

df<-merge(df1,df2,by=c("Entrez.ID","Gene.ID"),all=TRUE)

---

我如何告诉R在Entrez.ID变量匹配的地方，我希望Ensembl.ID匹配（即在可用的情况下用Ensembl.ID替换NA）？

我们可以使用

NA.locf

from

zoo

library(zoo)
df$Ensembl.ID <- with(df, ave(Ensembl.ID, Entrez.ID, FUN = function(x)
        na.locf(na.locf(x, na.rm = FALSE), fromLast = TRUE)))
df$Ensembl.ID
#[1] "ENSG00000113552" "ENSG00000113552" "ENSG00000249915" 
#[4] "ENSG00000249915" "ENSG00000095370"
#[6] "ENSG00000095370"

数据

---

df如果df1的Gene.ID和df2的Gene.ID是唯一的（df1中的ID不是df2中的ID），您可以简单地合并

df <- merge(df1,df2,by=c("Entrez.ID"),all.x=TRUE)

df我尝试从dplyr解决方案中获得完整的加入，它成功了-谢谢！
library(zoo)
df$Ensembl.ID <- with(df, ave(Ensembl.ID, Entrez.ID, FUN = function(x)
        na.locf(na.locf(x, na.rm = FALSE), fromLast = TRUE)))
df$Ensembl.ID
#[1] "ENSG00000113552" "ENSG00000113552" "ENSG00000249915" 
#[4] "ENSG00000249915" "ENSG00000095370"
#[6] "ENSG00000095370"

library(tidyverse)
full_join(df1, df2, by = c("Entrez.ID","Gene.ID")) %>%
    group_by(Entrez.ID) %>%
    fill(Ensembl.ID, .direction = 'up') %>%
    fill(Ensembl.ID, .direction = 'down')
# A tibble: 6 x 3
# Groups:   Entrez.ID [3]
#  Entrez.ID Gene.ID Ensembl.ID     
#      <int> <chr>   <chr>          
#1     10007 GNPDA1  ENSG00000113552
#2     10007 GPI     ENSG00000113552
#3     10016 ALG2    ENSG00000249915
#4     10016 PDCD6   ENSG00000249915
#5     10044 SH2D3C  ENSG00000095370
#6     10044 CHAT    ENSG00000095370

df <- structure(list(Entrez.ID = c(10007L, 10007L, 10016L, 10016L, 
10044L, 10044L), Gene.ID = c("GNPDA1", "GPI", "ALG2", "PDCD6", 
"SH2D3C", "CHAT"), Ensembl.ID = c(NA, "ENSG00000113552", NA, 
"ENSG00000249915", NA, "ENSG00000095370")), class = "data.frame", 
 row.names = c(NA, -6L))

df <- merge(df1,df2,by=c("Entrez.ID"),all.x=TRUE)

df <- rbind(df, df2)