R语言：如何在指定分组内将其他列值填充至空值单元格

Alright, let's sort out this issue you're facing—your original na.locf() approach wasn't working because it was operating on the entire dataset instead of within each Name group. That's a common pitfall when dealing with grouped data! Here are two reliable solutions to fill missing Int.Index values within each user group:

Option 1: Use dplyr + zoo::na.locf() (Leveraging the Function You Already Know) #

First, we'll group the data by Name so the fill operation happens independently for each person, then apply na.locf() to carry forward non-missing values.

library(dplyr)
library(zoo)

# Load your sample data
df <- read.table(header = T,text = 'Int.Index Int.Line Name Index Index.Line Amount
340 1 John 440 1 700
NA 2 John 440 2 100
NA 3 John 400 3 50
450 1 May 500 1 2401
5694 1 Paul 6023 1 1312
NA 2 Paul 6023 2 7244
NA 1 Tom 7889 1 221
NA 2 Tom 7889 2 2114
NA 1 Joe 231 1 1349
3455 1 Ben 3330 1 1353
NA 2 Ben 3330 2 8')

# Fill missing Int.Index values within each Name group
df_filled <- df %>%
  group_by(Name) %>%
  mutate(Int.Index = na.locf(Int.Index, na.rm = FALSE)) %>%  # Keep leading NAs (like Tom/Joe's groups)
  ungroup()

# View the result
print(df_filled)

The na.rm = FALSE argument ensures that if a group starts with an NA (like Tom or Joe's entries), those NAs stay intact instead of being filled with values from other groups.

Option 2: Use dplyr::fill() (A Cleaner, Native Alternative) #

If you'd prefer not to rely on the zoo package, dplyr has a built-in fill() function that does exactly what we need—no extra dependencies required:

library(dplyr)

df_filled <- df %>%
  group_by(Name) %>%
  fill(Int.Index, .direction = "down") %>%  # "down" means carry values forward to fill NAs below
  ungroup()

print(df_filled)

This method is more straightforward and achieves the same result as the first option.

Why Your Original Approach Failed #

When you used na.locf() without grouping, it treated the entire dataset as a single sequence. For example, the NA values in John's group would have been filled with May's Int.Index value (450) instead of John's initial 340. By grouping first, we restrict the fill operation to each individual's data, which is exactly what you need.

内容的提问来源于stack exchange，提问作者fabarin