根据数据集的条件设置颜色和框（箱线图）

Question

所以我有一个名为 Carbon 的数据集（Carbon$Graphite 确实包含一些在头部没有看到的值）：

       Mesa  TIC   TOC Graphite    TC
Kx V17   Ja 5.26 15.64       NA 20.90
Mu V17  Nej 4.08 11.32       NA 15.40
Ob V17  Nej 5.22 12.68       NA 17.90
Vä V17   Ja 6.45  6.35       NA 12.80
Ös V17  Nej 3.90  2.54       NA  6.44
Ig V17   Ja 8.20  3.20       NA 11.40

我想要一个显示 8 个盒子的箱线图，一个盒子只包含满足 Carbon$TIC[Carbon$Mesa=="Ja, ] 的值，一个等于 Carbon$TIC[Carbon$Mesa=="Ja, ]，一个等于 Carbon$TOC[Carbon$Mesa=="Ja, ] 等等。盒子的颜色（填充）是 Carbon$Mesa 所以“Ja”=红色盒子和“Nej”=蓝色盒子。我已经设法在不使用 ggpplot 的情况下做到了这一点，但我需要用 ggplot 来做到这一点（所以我所有的图表看起来都一样，其余的图表都是用 ggplot 完成的）。

我用来在没有 ggplot 的情况下制作的代码（这是我想要的，但与 ggplot 一样，旁边有图例）：

MesaJa <-Carbon[Carbon$Mesa=="Ja", ]
MesaNej <-Carbon[Carbon$Mesa=="Nej", ]

col.box<- c( rep("red", 3),   rep("blue", 3))

boxplot( list(MesaJa [, "TIC"], MesaJa [, "TOC"], MesaJa [, "TC"], 
              MesaNej[, "TIC"], MesaNej[, "TOC"], MesaNej[, "TC"] ),
         names=c("TIC", "TOC", "TC","TIC", "TOC", "TC") , 
         col=col.box
)  
legend("topleft", legend= c("Lime mud", "No lime mud"), pch=19, col=c("red","blue"), cex=0.7)

我尝试了几种不同的方法，但仍然无法正常工作。我得到的最接近的是：

Carbon$TIC_Ja <- ifelse(Carbon$Mesa=="Ja",Carbon$TIC, NA)
Carbon$TIC_Nej <- ifelse(Carbon$Mesa=="Nej",Carbon$TIC, NA)
Carbon$TOC_Ja <- ifelse(Carbon$Mesa=="Ja",Carbon$TOC, NA)
Carbon$TOC_Nej <- ifelse(Carbon$Mesa=="Nej",Carbon$TOC, NA)
Carbon$TC_Ja <- ifelse(Carbon$Mesa=="Ja",Carbon$TC, NA)
Carbon$TC_Nej <- ifelse(Carbon$Mesa=="Nej",Carbon$TC, NA)

Carbon.plot<-Carbon[ , c(1, 6:11)]
Carbon.key <- colnames(Carbon)  

ggplot(
  gather(Carbon.plot, key=Carbon.key, value="value", -"Mesa"),
  aes(x=factor(Carbon.key), y=as.numeric(value), fill= Carbon.key) 
  ) + 
  geom_boxplot() +
  scale_fill_manual(values=c("red", "blue", "red", "blue", "red", "blue"),
                    labels=c("Lime mud added", "No lime mud")
                    )

但这仍然不好，因为我只想要图例中的前两个帖子并想删除“NA”。而且我认为必须有一种更简单的方法，不涉及使用 ifelse 对数据框进行排序。我已经搜索过，但到目前为止还没有看到任何类似的例子。所以，请帮忙？

编辑：添加了有关 df 和会话信息的信息。但是在发布问题后，我开始更新一些软件包，但是事情并不顺利，所以目前我什至没有 ggplot 工作。

> str(Carbon)
'data.frame':   70 obs. of  5 variables:
 $ Mesa    : chr  "Ja" "Nej" "Nej" "Ja" ...
 $ TIC     : num  5.26 4.08 5.22 6.45 3.9 ...
 $ TOC     : num  15.64 11.32 12.68 6.35 2.54 ...
 $ Graphite: num  NA NA NA NA NA NA NA NA NA NA ...
 $ TC      : num  20.9 15.4 17.9 12.8 6.44 11.4 12.9 21.6 11.8 15.3 ...

> dput(Carbon)
structure(list(Mesa = c("Ja", "Nej", "Nej", "Ja", "Nej", "Ja", 
"Ja", "Ja", "Ja", "Ja", "Nej", "Nej", "Nej", "Ja", "Ja", "Nej", 
"Nej", "Nej", "Nej", "Ja", "Ja", "Ja", "Ja", "Nej", "Nej", "Nej", 
"Ja", "Ja", "Ja", "Nej", "Nej", "Nej", "Nej", "Ja", "Ja", "Ja", 
"Ja", "Nej", "Nej", "Nej", "Ja", "Ja", "Ja", "Nej", "Nej", "Nej", 
"Ja", "Nej", "Ja", "Ja", "Ja", "Ja", "Nej", "Nej", "Nej", "Ja", 
"Ja", "Ja", "Nej", "Nej", "Nej", "Nej", "Ja", "Ja", "Ja", "Ja", 
"Nej", "Nej", "Nej", "Ja"), TIC = c(5.26, 4.08, 5.22, 6.45, 3.9, 
8.2, 10.67, 7.43, 9.55, 8.19, 7.83, 4.04, 2.66, 4.93, 7.41, 3.25, 
4.47, 4.385, 3.48, 8.01, 9.49, 8.93, 6.03, 7.32, 3.84, 2.42, 
5.01, 3.87, 7, 4.8, 5.64, 5.76, 5.69, 8.7, 10.2, 9.78, 6.1, 8.07, 
4.33, 3.98, 7.39, 9.68, 9.67, 3.75, 5.07, 4.7, 4.86, 2.98, 8.05, 
10.29, 9.99, 6.65, 8.85, 4.82, 3.84, 8.49, 3.86, 6.63, 3.49, 
3.01, 4.83, 3.78, 8.95, 10.1, 8.15, 6.16, 8.15, 4.27, 3.96, 4.96
), TOC = c(15.64, 11.32, 12.68, 6.35, 2.54, 3.2, 2.23, 14.17, 
2.25, 7.11, 2.37, 17.16, 36.14, 13.47, 5.29, 17.95, 14.63, 3.85, 
6.31, 3.19, 2.81, 7.27, 23.07, 1.94, 26.19, 36.63, 23.19, 11.37, 
5.1, 2.39, 18.46, 2.17, 2.45, 1.7, 4.2, 4.92, 20.2, 0.86, 20.67, 
33.32, 5.11, 2.01, 0.53, 6.48, 29.51, 2.5, 9.41, 3.42, 3.04, 
4, 4.1, 11.15, 1.94, 20.66, 31.73, 1.81, 16.39, 20.75, 6.61, 
33.98, 2.48, 3.15, 1.65, 2.4, 12.63, 17.33, 0.99, 25.62, 38.63, 
17.93), Graphite = c(NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 
NA, NA, NA, NA, NA, NA, NA, NA, 0.13, NA, NA, NA, NA, NA, 0.07, 
0.05, NA, 0.06, NA, 0.03, NA, NA, NA, NA, NA, NA, NA, NA, NA, 
NA, NA, NA, NA, 0.07, 0.02, 0.08, 0.03, 0.33, NA, NA, NA, NA, 
NA, 0.02, 0.13, NA, 0.05, 0.02, NA, NA, NA, NA, NA, NA, NA, NA, 
NA, NA, NA, NA), TC = c(20.9, 15.4, 17.9, 12.8, 6.44, 11.4, 12.9, 
21.6, 11.8, 15.3, 10.2, 21.2, 38.8, 18.4, 12.7, 21.2, 19.1, 8.235, 
9.92, 11.2, 12.3, 16.2, 29.1, 9.26, 30.1, 39.1, 28.2, 15.3, 12.1, 
7.22, 24.1, 7.93, 8.14, 10.4, 14.4, 14.7, 26.3, 8.93, 25, 37.3, 
12.5, 11.7, 10.2, 10.3, 34.6, 7.28, 14.3, 6.73, 11.1, 14.3, 14.1, 
17.8, 10.8, 25.5, 35.7, 10.3, 20.3, 27.4, 10.1, 37, 7.33, 6.93, 
10.6, 12.5, 20.8, 23.5, 9.14, 29.9, 42.6, 22.9)), class = "data.frame", row.names = c("Kx V17", 
"Mu V17", "Ob V17", "Vä V17", "Ös V17", "Ig V17", "Va V17", "Gä V17", 
"Sk V17", "Fr V17", "Gr V17", "Bi V17", "As V17", "Kx H17", "Pi H17", 
"Mu H17", "Ob H17", "Do H17", "Ös H17", "Ig H17", "Va H17", "Gä H17", 
"Fr H17", "Gr H17", "Bi H17", "As H17", "So H17", "Kx V18", "Pi V18", 
"Mu V18", "Ob V18", "Do V18", "Ös V18", "Ig V18", "Va V18", "Gä V18", 
"Fr V18", "Gr V18", "Bi V18", "As V18", "So V18", "Kx H18", "Pi H18", 
"Mu H18", "Ob H18", "Do H18", "Vä H18", "Ös H18", "Ig H18", "Va H18", 
"Gä H18", "Fr H18", "Gr H18", "Bi H18", "As H18", "So H18", "Kx V19", 
"Pi V19", "Mu V19", "Ob V19", "Do V19", "Ös V19", "Ig V19", "Va V19", 
"Gä V19", "Fr V19", "Gr V19", "Bi V19", "As V19", "So V19"))

> sessionInfo()
R version 3.5.3 (2019-03-11)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18362)

Matrix products: default

locale:
[1] LC_COLLATE=Swedish_Sweden.1252  LC_CTYPE=Swedish_Sweden.1252   
[3] LC_MONETARY=Swedish_Sweden.1252 LC_NUMERIC=C                   
[5] LC_TIME=Swedish_Sweden.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
 [1] crayon_1.3.4     grid_3.5.3       R6_2.4.1         lifecycle_0.2.0  gtable_0.3.0    
 [6] magrittr_1.5     scales_1.1.0     rlang_0.4.5      rstudioapi_0.11  ellipsis_0.3.0  
[11] tools_3.5.3      glue_1.4.0       munsell_0.5.0    compiler_3.5.3   colorspace_1.4-1

Answer 1

这里一种可能的方法是使用例如tidyr 中的pivot_longer 函数（pivot_longer 在tidyr 1.0.0 中可用）将您的数据框转换为更长的格式：

library(tidyr)
library(dplyr)

Carbon %>% pivot_longer(cols = c(TIC, TOC, TC), names_to = "var",values_to = "val") %>%
  mutate(var = factor(var,levels = c("TIC","TOC","TC"))) 

# A tibble: 210 x 4
   Mesa  Graphite var     val
   <chr>    <dbl> <fct> <dbl>
 1 Ja          NA TIC    5.26
 2 Ja          NA TOC   15.6 
 3 Ja          NA TC    20.9 
 4 Nej         NA TIC    4.08
 5 Nej         NA TOC   11.3 
 6 Nej         NA TC    15.4 
 7 Nej         NA TIC    5.22
 8 Nej         NA TOC   12.7 
 9 Nej         NA TC    17.9 
10 Ja          NA TIC    6.45
# … with 200 more rows

然后，您可以使用interaction 将“Mesa”和“var”分组为x 值，并在ggplot2 中得到以下boxplot：

library(tidyr)
library(dplyr)
library(ggplot)

Carbon %>% pivot_longer(cols = c(TIC, TOC, TC), names_to = "var",values_to = "val") %>%
  mutate(var = factor(var,levels = c("TIC","TOC","TC"))) %>%
  ggplot(aes(x = interaction(var, Mesa), y = val, fill = Mesa))+
  geom_boxplot()+
  scale_x_discrete(labels = rep(c("TIC","TOC","TC"),2))+
  scale_fill_manual(values = c("red","blue"), labels = c("Lime mud", "No lime mud"))

它回答了你的问题吗？