如何用dplyr按ICD_Grouping分组统计Immunohistochemistry各水平频数

Hey there! Let's get your dplyr code sorted to produce that wide-format data frame you need. Here's a step-by-step breakdown with working code:

Step 1: Fix the Grouping & Counting #

Your initial code only grouped by ICD_Grouping, but we need to group by both ICD_Grouping and Immunohistochemistry to capture counts for each factor level (Yes/No/N/A) within each ICD group. We can use dplyr::count() as a handy shortcut for group_by() + summarise(n()):

Step 2: Reshape to Wide Format #

Instead of struggling with spread() (which is now soft-deprecated in the tidyverse), we'll use tidyr::pivot_wider()—it's more intuitive and flexible. If you still want to use spread(), I'll include that version too.

Recommended: Using pivot_wider()

library(dplyr)
library(tidyr)

null.ta <- dbdata %>%
  filter(MutGroup == "Null") %>%
  # Count occurrences of each Immunohistochemistry level per ICD_Grouping
  count(ICD_Grouping, Immunohistochemistry) %>%
  # Reshape to wide format, filling missing levels with 0
  pivot_wider(
    names_from = Immunohistochemistry,  # Which column becomes new column names
    values_from = n,                    # Which column provides the values
    values_fill = 0                     # Fill empty cells with 0
  )

Legacy: Using spread()

If you prefer to stick with spread():

null.ta <- dbdata %>%
  filter(MutGroup == "Null") %>%
  group_by(ICD_Grouping, Immunohistochemistry) %>%
  summarise(count = n(), .groups = "drop") %>%  # Drop grouping after summarizing
  spread(
    key = Immunohistochemistry,
    value = count,
    fill = 0
  )

Step 3: Clean Up Unwanted Columns #

If your data has extra factor levels (like that 4th level you mentioned), just use select() to keep only the columns you need. Note that N/A has special characters, so we wrap it in backticks:

null.ta <- null.ta %>%
  select(ICD_Grouping, Yes, No, `N/A`)

Final Output #

This will give you exactly the format you're looking for:

内容的提问来源于stack exchange，提问作者Seb Walpole