生成PPT时自定义插入绘图尺寸的技术咨询

Hey there, I totally get your frustration—trying to size ggplot and editable rvg plots directly in PowerPoint via officer can feel like fighting with hidden settings. Let’s break down solutions for both plot types that fix the scaling issues you ran into:

1. For Regular ggplot Objects: Use ph_location() for Precise Control #

Forget relying on ph_location_type for sizing ggplot outputs. Instead, use ph_location() to define exact width, height, and position (optional) directly. This ensures your plot scales uniformly, with labels and graphics staying in proper proportion:

library(flextable)
library(rvg)
library(officer)
library(ggplot2)
path_out <- "."

# Prep your ggplot
p1 <- iris %>% ggplot() + 
  geom_point(aes(Sepal.Length, Petal.Length, color = Species), size = 3) + 
  theme_minimal()

my_pres <- read_pptx() %>% 
  # Slide 1: Explicitly set plot dimensions
  add_slide(layout = "Title and Content", master = "Office Theme") %>% 
  ph_with(
    value = p1, 
    location = ph_location(
      left = 1, top = 1, # Optional: Adjust position on slide
      width = 8, height = 5 # Set your desired size (in inches)
    )
  ) %>% 
  print(target = file.path(path_out,"example_ggplot.pptx"))

Dimensions here use inches, which align with PowerPoint’s default slide size (13.33x7.5 inches), so you can easily map plot size to your slide layout.

2. For Editable rvg Plots: Adjust Canvas Size First #

The dml() output is vector-based, so width/height in ph_with() won’t work directly. Instead, set the canvas size in dml() itself, then tell PowerPoint to respect that size:

Method A: Preset Canvas Size in dml() #

Define your desired dimensions when creating the rvg object, then enable use_loc_size = TRUE to apply it in the slide:

# Create editable rvg plot with fixed canvas size
p2 <- dml(
  ggobj = p1,
  width = 6, height = 6 # Set size here (in inches)
)

my_pres <- read_pptx() %>% 
  add_slide() %>% 
  ph_with(
    value = p2, 
    location = ph_location_type("body"),
    use_loc_size = TRUE # Tell PowerPoint to use the rvg's preset size
  ) %>% 
  print(target = file.path(path_out,"example_rvg.pptx"))

This keeps your editable plot’s elements (labels, points, lines) scaled perfectly together.

Method B: Fit to Placeholder Size #

If you want the rvg plot to fill the entire body placeholder, use ph_location_type with use_loc_size = TRUE to auto-adapt:

my_pres <- read_pptx() %>% 
  add_slide() %>% 
  ph_with(
    value = p2, 
    location = ph_location_type("body"),
    use_loc_size = TRUE
  ) %>% 
  print(target = file.path(path_out,"example_rvg_fill.pptx"))

Why Your Previous Attempts Had Issues #

• use_loc_size = FALSE works for images because they have fixed pixel dimensions, but rvg objects are vector-based—this parameter actually needs to be TRUE to force PowerPoint to respect the size you define for the rvg object.
• Adjusting pixel resolution (like dpi) messes up scaling because ggplot’s label sizes are tied to font settings (not pixels), leading to the mismatched label/plot scaling you saw.

内容的提问来源于stack exchange，提问作者Ian Quest