ggplot theme_bw不适用于读取数据，但相同的确切代码对mtcars数据也可以正常工作

如何解决ggplot theme_bw不适用于读取数据，但相同的确切代码对mtcars数据也可以正常工作

我有如下数据，其中约2000行其中sample_name是唯一的（它实际上是一个因子值，但我在这里进行了转换以使其隐藏）：

head(df)
  sample_name metric metric_name metric_mean metric_sd year_month
1           1      0   my_metric    154260.5  76782.54    2020-08
2           2 156845   my_metric    154260.5  76782.54    2020-08
3           3  78015   my_metric    154260.5  76782.54    2020-08
4           4  87620   my_metric    154260.5  76782.54    2020-08
5           5 102581   my_metric    154260.5  76782.54    2020-08
6           6 145976   my_metric    154260.5  76782.54    2020-08

我正在使用RMD内容来生成带有图的PDF。如使用rmarkdown PDF转换生成的输出图所示，生成的图将删除theme_bw()。

---
title: "Example"
author: John Doe
output: pdf_document
---

# Here is my plot

This is a plot of mpg vs. number of cylinders in a car.

```{r,results = 'asis',echo = FALSE,eval = TRUE,message = FALSE,warn = FALSE}
  library(ggplot2)
  library(data.table)

  df <- fread('/tmp/gopala.csv',data.table = FALSE)
    metric_mean <- unique(df$metric_mean)
    metric_sd <- unique(df$metric_sd)
    metric_name <- unique(df$metric_name)
    year_month <- unique(df$year_month)

  ggplot(df,aes(x = sample_name,y = metric)) +
    geom_point(size = 0.5,color = '#1F77B4') +
    labs(
      title = paste0('CHART FOR PERIOD ',year_month),x = 'NAME',y = toupper(metric_name)
      ) +
    geom_hline(yintercept = metric_mean,color = 'red',size = 0.5) +
    geom_hline(
      yintercept = metric_mean + metric_sd,color = 'blue',size = 0.5
      ) +
    geom_hline(
      yintercept = metric_mean - metric_sd,linetype = 'dashed',size = 0.5
      ) +
    geom_hline(
      yintercept = metric_mean + 2 * metric_sd,color = 'purple',size = 0.5
      ) +
    geom_hline(
      yintercept = metric_mean - 2 * metric_sd,size = 0.5
      ) +
    geom_hline(
      yintercept = metric_mean + 3 * metric_sd,color = 'green',size = 0.5
      ) +
    geom_hline(
      yintercept = metric_mean - 3 * metric_sd,size = 0.5
      ) +
    geom_text(
      aes(
        label = paste0('Mean (',round(metric_mean,2),')'),x = Inf,y = metric_mean,fontface = 'bold'
        ),hjust = 1
      ) +
    geom_text(
      aes(
        label = paste0('+1SD (',round(metric_mean + metric_sd,y = metric_mean + metric_sd,color = 'blue'
        ),hjust = 1
      ) +
    geom_text(
      aes(
        label = paste0('-1SD (',round(metric_mean - metric_sd,y = metric_mean - metric_sd,hjust = 1
      ) +
    geom_text(
      aes(
        label = paste0('+2SD (',round(metric_mean + 2 * metric_sd,y = metric_mean + 2 * metric_sd,color = 'purple'
        ),hjust = 1
      ) +
    geom_text(
      aes(
        label = paste0('-2SD (',round(metric_mean - 2 * metric_sd,y = metric_mean - 2 * metric_sd,hjust = 1
      ) +
    geom_text(
      aes(
        label = paste0('+3SD (',round(metric_mean + 3 * metric_sd,y = metric_mean + 3 * metric_sd,color = 'green'
        ),hjust = 1
      ) +
    geom_text(
      aes(
        label = paste0('-3SD (',round(metric_mean - 3 * metric_sd,y = metric_mean - 3 * metric_sd,hjust = 1
      ) +
    theme_bw() +
    theme(
      plot.title = element_text(hjust = 0.5),axis.text.x = element_blank(),axis.ticks.x = element_blank(),legend.position = 'none'
      )
```

将前几行更改为以下内容，可使一切正常：

  metric_mean <- mean(mtcars$mpg)
  metric_sd <- sd(mtcars$mpg)

  ggplot(mtcars,aes(x = wt,y = mpg)) +
    geom_point(size = 0.5,color = '#1F77B4') +
    labs(
      title = 'WT vs. MPG',x = 'WT',y = 'MPG'
      ) +

对于我的一生，我无法弄清原因。

附加：如果有人知道如何在不使用绝对值的情况下将我的“文本”标签微调（相对比例），那也将很有帮助。