这个问题在这里已经有了答案:
Nested facets in ggplot2 spanning groups
(2 个回答)
去年关闭。
我在 ggplot 中创建了一个图表里面有两个变量 facet_grid .
我希望每个方面的标题仅在该方面的中心重复一次。
例如，第一个原始(上面)中的 0 和 1 将只出现一次并且出现在中间。
在我原来的情节中，每个方面的情节数量不相等。因此，使用 patchwork 将两个图拼凑在一起/cowplot/ggpubr效果不佳。
我更喜欢只使用 ggplot 的解决方案/hack .
样本数据:
df <- head(mtcars, 5)
示例图:
df %>%
ggplot(aes(gear, disp)) +
geom_bar(stat = "identity") +
facet_grid(~am + carb,
space = "free_x",
scales = "free_x") +
ggplot2::theme(
panel.spacing.x = unit(0,"cm"),
axis.ticks.length=unit(.25, "cm"),
strip.placement = "outside",
legend.position = "top",
legend.justification = "center",
legend.direction = "horizontal",
legend.key.size = ggplot2::unit(1.5, "lines"),
# switch off the rectangle around symbols
legend.key = ggplot2::element_blank(),
legend.key.width = grid::unit(2, "lines"),
# # facet titles
strip.background = ggplot2::element_rect(
colour = "black",
fill = "white"),
panel.background = ggplot2::element_rect(
colour = "white",
fill = "white"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank())
编辑 - 新数据
我创建了一个示例数据，它更准确地类似于我的实际数据。
structure(list(par = c("Par1", "Par1", "Par1", "Par1", "Par1",
"Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1",
"Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1",
"Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par2", "Par2",
"Par2"), channel_1 = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 11L, 11L, 11L, 11L,
11L, 11L, 11L, 11L, 11L, 1L, 1L, 1L), .Label = c("Center", "Left \nFrontal",
"Left \nFrontal Central", "Left \nCentral Parietal", "Left \nParietal Ooccipital",
"Left", "Right \nFrontal", "Right \nFrontal Central", "Right \nCentral Parietal",
"Right \nParietal Ooccipital", "Right"), class = "factor"), freq = structure(c(1L,
1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L,
3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("Alpha",
"Beta", "Gamma"), class = "factor"), group = c("a", "b", "c",
"a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a",
"b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b",
"c"), m = c(0.488630500442935, 0.548666228768508, 0.0441536349332613,
0.304475866391531, 0.330039488441422, 0.0980622573307064, 0.0963996979198171,
0.301679466108907, 0.240618782227119, 0.35779695722622, 0.156116647839907,
0.0274546218676152, 0.0752501569920047, 0.289342864254614, 0.770518960576786,
0.548130676907356, 0.180158614358946, 0.238520826021687, 0.406326198917495,
0.159739769132509, 0.140739952534666, 0.295427640977557, 0.106130817023844,
0.214006898241167, 0.31081727835652, 0.366982521446529, 0.264432086988446,
0.0761271112139142, 0.0811642772125171, 0.0700455890939194),
se = c(0.00919040825504951, 0.00664655073810519, 0.0095517721611042,
0.00657090455386036, 0.00451135146762504, 0.0188625074573698,
0.00875378313351897, 0.000569521129673224, 0.00691447732630984,
0.000241814142091401, 0.0124584589176995, 0.00366855139256551,
0.0072981677277562, 0.0160663614099261, 0.00359337442316408,
0.00919725279757502, 0.040856967817406, 0.00240910563984416,
0.0152236046767608, 0.00765487375180611, 0.00354140237391633,
0.00145468584619171, 0.0185141245423404, 0.000833307847848054,
0.0038193622895167, 0.0206130436440409, 0.0066911922721337,
7.3079999953491e-05, 0.0246233416039572, 0.00328150956514463
)), row.names = c(NA, -30L), class = c("tbl_df", "tbl", "data.frame"
))
阴谋:
df %>%
ggplot(aes(channel_1, m,
group = group,
fill = group,
color = group)) +
facet_grid(~par + freq,
space="free_x",
scales="free_x") +
geom_errorbar(
aes(min = m - se, ymax = m + se, alpha = 0.01),
width = 0.2, size = 2, color = "black",
position = position_dodge(width = 0.6)) +
geom_bar(stat = "identity",
position = position_dodge(width = 0.6),
# color = "black",
# fill = "white",
width = 0.6,
size = 2, aes(alpha = 0.01)) +
scale_shape_manual(values = c(1, 8, 5)) +
labs(
color = "",
fill = "",
shape = "") +
guides(
color = FALSE,
shape = FALSE) +
scale_alpha(guide = "none")
最佳答案
最快的黑客:用情节伪造方面并结合。这需要一些黑客攻击，但它可能仍然比搞乱 grobs 更少黑客攻击:
library(patchwork)
library(tidyverse)
df <- head(mtcars,5)
df <- df %>% mutate(am_carb = factor(paste(am,carb,sep = '_'),
labels = c( ' 1','2','1','4')))
##note!! the blank space in ' 1' label is on purpose!!! this is to make those labels unique, otherwise it would consider both '1' the same category!!
p1 <-
df %>%
ggplot(aes(gear, disp)) +
geom_bar(stat = "identity") +
facet_grid(~am_carb, scales = "free_x") +
theme(panel.spacing.x = unit(0,"cm"),
plot.margin = margin(t = -2),
strip.background = element_rect(colour = "black",fill = "white"),
panel.background = element_rect(colour = "white", fill = "white"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank())
p2 <-
df %>%
ggplot(aes(gear, disp)) +
geom_blank() +
facet_grid(~ am, scales = "free_x") +
theme(panel.spacing.x = unit(0,"cm"),
axis.text = element_blank(),
axis.ticks = element_blank(),
axis.title = element_blank(),
plot.margin = margin(b = -2),
strip.background = element_rect(colour = "black",fill = "white"),
panel.background = element_rect(colour = "white", fill = "white"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank())
p2/p1 + plot_layout(heights = c(0.1,100) )
创建于 2020-03-24 由 reprex package (v0.3.0)
用新数据更新 - 一些更复杂的方面。事实上，拼凑在这里很困难。在将假面转换为网格对象并更改宽度后，更容易将假面与牛图结合起来。内所有
cowplot .mydat <- structure(list(par = c("Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par1", "Par2", "Par2", "Par2"), channel_1 = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 11L, 11L, 11L, 11L, 11L, 11L, 11L, 11L, 11L, 1L, 1L, 1L), .Label = c("Center", "Left \nFrontal", "Left \nFrontal Central", "Left \nCentral Parietal", "Left \nParietal Ooccipital", "Left", "Right \nFrontal", "Right \nFrontal Central", "Right \nCentral Parietal", "Right \nParietal Ooccipital", "Right"), class = "factor"), freq = structure(c(1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("Alpha", "Beta", "Gamma"), class = "factor"), group = c("a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c", "a", "b", "c"), m = c(0.488630500442935, 0.548666228768508, 0.0441536349332613, 0.304475866391531, 0.330039488441422, 0.0980622573307064, 0.0963996979198171, 0.301679466108907, 0.240618782227119, 0.35779695722622, 0.156116647839907, 0.0274546218676152, 0.0752501569920047, 0.289342864254614, 0.770518960576786, 0.548130676907356, 0.180158614358946, 0.238520826021687, 0.406326198917495, 0.159739769132509, 0.140739952534666, 0.295427640977557, 0.106130817023844, 0.214006898241167, 0.31081727835652, 0.366982521446529, 0.264432086988446, 0.0761271112139142, 0.0811642772125171, 0.0700455890939194), se = c(0.00919040825504951, 0.00664655073810519, 0.0095517721611042, 0.00657090455386036, 0.00451135146762504, 0.0188625074573698, 0.00875378313351897, 0.000569521129673224, 0.00691447732630984, 0.000241814142091401, 0.0124584589176995, 0.00366855139256551, 0.0072981677277562, 0.0160663614099261, 0.00359337442316408, 0.00919725279757502, 0.040856967817406, 0.00240910563984416, 0.0152236046767608, 0.00765487375180611, 0.00354140237391633, 0.00145468584619171, 0.0185141245423404, 0.000833307847848054, 0.0038193622895167, 0.0206130436440409, 0.0066911922721337, 7.3079999953491e-05, 0.0246233416039572, 0.00328150956514463)), row.names = c(NA, -30L), class = c("tbl_df", "tbl", "data.frame"))
library(tidyverse)
library(cowplot)
#>
#> ********************************************************
#> Note: As of version 1.0.0, cowplot does not change the
#> default ggplot2 theme anymore. To recover the previous
#> behavior, execute:
#> theme_set(theme_cowplot())
#> ********************************************************
mydat <- mydat %>% mutate(par_freq = factor(paste(par,freq,sep = '_'), labels = c('Alpha', 'Beta', 'Gamma', 'Gamma ' )))
p1 <-
mydat %>%
ggplot(aes(channel_1, m, group = group, fill = group, color = group)) +
geom_bar(stat = "identity") +
facet_grid( ~ par_freq, scales = "free_x", space="free_x") +
theme(panel.spacing.x = unit(0,"cm"),
plot.margin = margin(t = -2),
strip.background = element_rect(colour = "black",fill = "white"),
panel.background = element_rect(colour = "white", fill = "white"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
legend.position = 'none')
p2 <-
mydat %>%
ggplot(aes(channel_1, m, group = group, fill = group, color = group)) +
geom_blank() +
facet_grid(~ par) +
theme(panel.spacing.x = unit(0,"cm"),
axis.text = element_blank(),
axis.ticks = element_blank(),
axis.title = element_blank(),
plot.margin = margin(b = -2),
strip.background = element_rect(colour = "black",fill = "white"),
panel.background = element_rect(colour = "white", fill = "white"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank())
gt <- cowplot::as_gtable(p2)
gt$widths[5] <- 8*gt$widths[7]
cowplot::plot_grid(gt, p1, align = "v", axis = 'l',nrow = 2, rel_heights = c(5, 100))
# you need to play around with the values unfortunately.
创建于 2020-03-24 由 reprex package (v0.3.0)
一些额外的想法
我在想一个人无法绕过这样的黑客 - 因为原始图的 gtable_layout(带有两个方面变量)显示整个方面条是一个 grob! THis answer proved me wrong - the grob contains a nested table for both strips! .但由于
ggnomics，有一个更简单的解决方案。包 - 见我的第二个答案p_demo <- ggplot(mydat, aes(channel_1, m)) +
geom_bar(stat = "identity") +
facet_grid(~par +freq , space = "free_x", scales = "free_x") +
theme(panel.spacing.x = unit(0,"cm"))
gt <- cowplot::as_gtable(p_demo)
gtable::gtable_show_layout(gt)
创建于 2020-03-24 由 reprex package (v0.3.0)
关于r - ggplot - 集中 facet_grid 标题并且只出现一次，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/60822398/