按一列匹配两个数据帧，并从R中另一列中匹配的行中减去

我正在努力做到以下几点：

将R数据帧拆分为控件和rest（示例）

匹配一列控件和示例数据帧（单词匹配）

对于该列中的每个匹配项，减去另一列中的样本控制值

打印输出数据框中的所有匹配行和非匹配行

我试过：

#  input

My.Data <- structure(list(V1 = structure(1:9, .Label = c("a1", "a2", "a3", 
                                                         "a4", "a5", "control1", "control2", "control3", "control4"), class = "factor"), 
                          V2 = structure(c(1L, 1L, 2L, 3L, 5L, 1L, 2L, 3L, 4L), .Label = c("otu1", 
                                                                                           "otu2", "otu3", "otu4", "otu6"), class = "factor"), V3 = structure(c(4L, 
                                                                                                                                                                5L, 6L, 9L, 8L, 3L, 1L, 2L, 7L), .Label = c("ee", "tt", "w", 
                                                                                                                                                                                                            "xx", "xxx", "xy", "yy", "z44", "zz"), class = "factor"), 
                          V4 = c(44L, 52L, 11L, 22L, 91L, 4L, 34L, 33L, 11L)), class = "data.frame", row.names = c(NA, 
                                                                                                                   -9L))

# split groups

control<-My.Data[grep("^control*", My.Data$V1), ]
sample<-My.Data[!grepl("^control*",My.Data$V1),]

# match V2 in control and samples (example match: otu1 with otu1..)
?

# Whenever match found in V2 (multiple match is possible), subtract sample-control values in V4
?

# print all matched (and non-matched) rows in a dataframe

---

谢谢。

我们可以将

V2

上的两个数据集合并并减去列“V4”

library(data.table)
setDT(sample)[control, V4 := V4 - i.V4, on = .(V2)]
sample
#   V1   V2  V3  V4
#1: a1 otu1  xx  40
#2: a2 otu1 xxx  48
#3: a3 otu2  xy -23
#4: a4 otu3  zz -11
#5: a5 otu6 z44  91

如果我们想与“控件”的不匹配行绑定

rbind(sample, setDT(control)[!sample, on = .(V2)])
#         V1   V2  V3  V4
#1:       a1 otu1  xx  40
#2:       a2 otu1 xxx  48
#3:       a3 otu2  xy -23
#4:       a4 otu3  zz -11
#5:       a5 otu6 z44  91
#6: control4 otu4  yy  11

---

在

tidyverse

中，我们可以使用

left\u join

和

bind\u rows

与

anti\u join

library(dplyr)
left_join(sample, control %>% 
     select(V2, V4), by = 'V2') %>%
  transmute(V1, V2, V3, V4 = coalesce(V4.x-V4.y, V4.x)) %>% 
  bind_rows(anti_join(control, sample, by = 'V2'))

---

一个

dplyr

选项可以是：

My.Data %>%
 group_by(V2) %>%
 filter(n() > 1) %>%
 mutate(V4 = V4 - V4[grepl("^control", V1)]) %>%
 filter(!grepl("^control", V1)) %>%
 bind_rows(My.Data %>%
            group_by(V2) %>%
            filter(n() == 1))

  V1       V2    V3       V4
  <fct>    <fct> <fct> <int>
1 a1       otu1  xx       40
2 a2       otu1  xxx      48
3 a3       otu2  xy      -23
4 a4       otu3  zz      -11
5 a5       otu6  z44      91
6 control4 otu4  yy       11

My.Data%>%
分组依据（V2）%>%
过滤器（n（）>1）%>%
变异（V4=V4-V4[grepl（“^control”，V1）]）%>%
过滤器（！grepl（“^control”，V1））%>%
绑定_行（My.Data%>%
分组依据（V2）%>%
过滤器（n（）=1））
V1 V2 V3 V4
1 a1 otu1 xx 40
2 a2 otu1 xxx 48
3 a3 otu2 xy-23
4 a4 otu3 zz-11
5 a5 otu6 z44 91
6控制4 otu4 yy 11

My.Data %>%
 group_by(V2) %>%
 filter(n() > 1) %>%
 mutate(V4 = V4 - V4[grepl("^control", V1)]) %>%
 filter(!grepl("^control", V1)) %>%
 bind_rows(My.Data %>%
            group_by(V2) %>%
            filter(n() == 1))

  V1       V2    V3       V4
  <fct>    <fct> <fct> <int>
1 a1       otu1  xx       40
2 a2       otu1  xxx      48
3 a3       otu2  xy      -23
4 a4       otu3  zz      -11
5 a5       otu6  z44      91
6 control4 otu4  yy       11