针对特定分组修改dplyr中summarize()计算方法（附示例数据集）

First, let's complete your sample dataset with dummy count values so we can work through examples properly:

library(dplyr)

df <- structure(
  list(
    x = structure(
      c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 7L),
      .Label = c("1", "2", "3", "4", "5", "6", "Total"),
      class = "factor"
    ),
    y = structure(
      c(1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L),
      .Label = c("7", "8", "9", "Total"),
      class = "factor"
    ),
    z = structure(
      c(1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L),
      .Label = c("10", "11"),
      class = "factor"
    ),
    count = c(56, 32, 45, 29, 61, 18, 241, 119, 56, 32, 45, 29, 61, 18, 241, 119)
  ),
  class = "data.frame",
  row.names = c(NA, -16L)
)

Scenario 1: Summarize only non-"Total" groups #

If you want to compute aggregated values (like sum of count) for groups where neither x nor y is "Total", filter those rows first before grouping and summarizing:

non_total_summary <- df %>%
  filter(x != "Total", y != "Total") %>%
  group_by(x, y, z) %>%
  summarize(total_count = sum(count), .groups = "drop")

print(non_total_summary)

This gives you clean aggregated counts for every valid combination of x, y, and z excluding the precomputed total rows.

Scenario 2: Validate or recalculate "Total" groups #

If you need to ensure the existing "Total" rows in your dataset are accurate, you can compute these totals programmatically and replace the existing values:

# Calculate true totals for each z value
calculated_totals <- df %>%
  filter(x != "Total", y != "Total") %>%
  group_by(z) %>%
  summarize(calculated_total = sum(count), .groups = "drop") %>%
  mutate(x = "Total", y = "Total")

# Combine with original non-total rows to create an updated dataset
updated_df <- df %>%
  filter(x != "Total" | y != "Total") %>%
  bind_rows(calculated_totals) %>%
  arrange(x, y, z)

print(updated_df)

This replaces the existing "Total" rows with values derived directly from your raw data, ensuring consistency.

Scenario 3: Custom summary logic for specific subgroups #

If you want to apply different calculations to specific groups (e.g., sum for z="10" and average for z="11"), use conditional logic inside summarize():

custom_group_summary <- df %>%
  filter(x != "Total", y != "Total") %>%
  group_by(x, y) %>%
  summarize(
    z10_total = sum(count[z == "10"]),
    z11_avg = mean(count[z == "11"]),
    .groups = "drop"
  )

print(custom_group_summary)

This lets you tailor your summary exactly to the groups and metrics you care about.

If you had a specific grouping or calculation in mind that isn't covered here, feel free to share more details and I can adjust the code accordingly!

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