用ggplot2同时设置双y轴图的左右极限范围

Question

给出一个数据df

df <- structure(list(date = c("2022-4-30", "2022-5-5", "2022-5-6", 
"2022-5-9", "2022-5-10", "2022-5-11", "2022-5-12", "2022-5-13", 
"2022-5-16", "2022-5-17", "2022-5-18", "2022-5-19", "2022-5-20"
), value1 = c(6, 6.1, 6.5, 6.9, 7.3, 7.7, 8.1, 8, 7.9, 7.8, 7.7, 
7.6, 7.5), value2 = c(-1L, -2L, 0L, 6L, 7L, 5L, 3L, 10L, 18L, 
16L, 12L, 19L, 20L), value3 = c(2.7, 2.4, 1.7, 0.5, -0.5, 0.2, 
-0.3, -0.2, 0.4, 0.9, 1.3, 1.1, 1)), class = "data.frame", row.names = c(NA, 
-13L))

退出：

        date value1 value2 value3
1  2022-4-30    6.0     -1    2.7
2   2022-5-5    6.1     -2    2.4
3   2022-5-6    6.5      0    1.7
4   2022-5-9    6.9      6    0.5
5  2022-5-10    7.3      7   -0.5
6  2022-5-11    7.7      5    0.2
7  2022-5-12    8.1      3   -0.3
8  2022-5-13    8.0     10   -0.2
9  2022-5-16    7.9     18    0.4
10 2022-5-17    7.8     16    0.9
11 2022-5-18    7.7     12    1.3
12 2022-5-19    7.6     19    1.1
13 2022-5-20    7.5     20    1.0

我使用下面的代码绘制双y轴( value1和value3的左侧，value2的右侧)：

df_m <- melt(df, id.vars = 'date')
df_m <- df_m %>% 
  mutate(date=as.Date(date))

df_m1 <- df_m %>% 
  filter(variable %in% c("value1", 'value3')) 

df_m2 <- df_m %>% 
  filter(variable %in% c("value2")) %>%
  mutate(value = value * 0.6)


df_m2
coeff = 1/0.6

ggplot() +
  geom_line(data = df_m1[!is.na(df_m1$value), ], aes(x = date, y = value, col = variable), alpha = 1, size = 1) +
  geom_line(data = df_m2[!is.na(df_m2$value), ], aes(x = date, y = value, col = variable), alpha = 1, size = 1) +
  scale_y_continuous(
    name = '',
    # limits=c(-1, 9),
    sec.axis = sec_axis(~.*coeff, name = "")) 

退出：

​

​

现在，我想知道是否可以使用limits=c(-1, 9)同时设置左y轴和右y轴的limits=c(-2, 20)？

我尝试使用下面的代码，但右y轴被更改为limits=c(-2, 15)，而不是上面的图。类似的双y轴图的替代解决方案也将受到欢迎。

ggplot() +
  geom_line(data = df_m1[!is.na(df_m1$value), ], aes(x = date, y = value, col = variable), alpha = 1, size = 1) +
  geom_line(data = df_m2[!is.na(df_m2$value), ], aes(x = date, y = value, col = variable), alpha = 1, size = 1) +
  scale_y_continuous(
    name = '',
    limits=c(-1, 9),
    sec.axis = sec_axis(~.*coeff, name = ""))

​

​

参考资料：

https://r-graph-gallery.com/line-chart-dual-Y-axis-ggplot2.html

Answer 1

这是编写ggh4x::help_secondary()的原因之一。从本质上讲，它计算出了获得二次范围以匹配主范围和逆范围所需的转换。将该变换应用于二次数据，并将其逆作为二次轴的trans参数。

在您的示例中，我们可以删除mutate()位，因为这是在sec$proj()中处理的。

(免责声明:我写了ggh4x)

# df <- structure(...) # omitted for brevity

library(ggplot2)
library(dplyr)
library(reshape2)

df_m <- melt(df, id.vars = 'date')
df_m <- df_m %>% 
  mutate(date=as.Date(date))

df_m1 <- df_m %>% 
  filter(variable %in% c("value1", 'value3')) 

df_m2 <- df_m %>% 
  filter(variable %in% c("value2"))

sec <- ggh4x::help_secondary(
  name = "",
  primary = c(-1, 9), secondary = c(-2, 20),
)

ggplot() +
  geom_line(data = df_m1[!is.na(df_m1$value), ], 
            aes(x = date, y = value, col = variable), alpha = 1, size = 1) +
  geom_line(data = df_m2[!is.na(df_m2$value), ], 
            aes(x = date, y = sec$proj(value), col = variable), alpha = 1, size = 1) +
  scale_y_continuous(
    limits = c(-1, 9),
    name = '',
    sec.axis = sec) 

​

​