如何无需转换为长数据框，直接基于mtcars宽数据框使用ggplot2绘制facet_wrap直方图？

Great question! The core challenge here is that facet_wrap relies on a categorical variable to split your plots into facets, which wide-format data like the original mtcars doesn't have—each numeric variable lives in its own separate column.

While you can't directly fill in the blanks in your proposed code structure without restructuring the data, you can avoid creating a standalone long data frame with tidyr::pivot_longer by using base R's stack() function directly inside the ggplot() call. This converts your selected numeric columns to long format on the fly, no intermediate object required.

Here's the completed code that produces exactly the same facet histogram output as your original workflow:

ggplot(stack(mtcars[, c("mpg", "disp", "hp", "drat", "wt", "qsec")]), aes(x = values)) +
  facet_wrap(~ind, scales = 'free') +
  geom_histogram(bins = 32, color = 'black', fill = 'lightblue')

Quick breakdown: #

• stack(mtcars[, c("mpg", "disp", "hp", "drat", "wt", "qsec")]) transforms your chosen numeric columns into a two-column data frame:
  • values: Contains all the numeric values from the original columns
  • ind: Stores the original column names (this is the categorical variable we use for facetting)
• aes(x = values) maps the numeric values to the x-axis of each histogram
• facet_wrap(~ind, scales = 'free') creates a separate facet for each original variable, with free scales to handle the different value ranges (matching your original code's behavior)

If you’re set on avoiding any data reshaping functions at all (even base R ones), there’s no way to use facet_wrap directly—since it needs a grouping variable that doesn’t exist in the wide data. The alternative would be to create individual histograms for each variable and combine them with a package like patchwork, but that doesn’t leverage facet_wrap as you requested.

内容的提问来源于stack exchange，提问作者hachiko