如何在R中将特定格式文本文件转换为指定结构的DataFrame

Let's break down how to fix your issue and get the desired DataFrame structure. First, let's unpack what went wrong with your original code:

Why Your Code Threw an Error #

Your line separate_rows(circ, convert = TRUE) references circ as a column name, but when you read the CSV with read.csv(text=circ, sep=" ", header = FALSE), the resulting dataframe uses default column names like V1, V2, etc.—there's no column named circ. That's exactly why you hit the "Unknown column ASDDSDAS" error.

Step-by-Step Fix #

Here's a working approach to transform your input text into the target structure:

1. Read the input correctly: Your input is a single line of space-separated values, so we'll first read it into a wide dataframe.
2. Reshape to long format: Convert all columns into a single column of values to process each entry individually.
3. Distinguish circle IDs from node IDs: Circle IDs are non-numeric strings, while node IDs are numeric values.
4. Pair circles with their nodes: Use a fill operation to attach each circle ID to all subsequent nodes until the next circle ID appears.
5. Clean up the final dataframe: Filter out redundant rows and rename columns to match your target structure.

Working Code

library(tidyverse)

# Replace this with your actual file reading (e.g., read_file(circ_file))
circ_text <- "ASDDSDAS 111 132 456 DSSADSDA 1 12335"

# Read the space-separated line into a wide dataframe
cir_wide <- read.table(text = circ_text, sep = " ", header = FALSE, stringsAsFactors = FALSE)

# Reshape to long format and track row positions
cir_long <- cir_wide %>%
  pivot_longer(cols = everything(), names_to = NULL, values_to = "value") %>%
  mutate(row_pos = row_number())

# Mark which entries are circle IDs (non-numeric)
cir_long <- cir_long %>%
  mutate(is_circle = !str_detect(value, "^\\d+$"))

# Fill circle IDs down to their corresponding nodes, then clean up
cir_final <- cir_long %>%
  mutate(circleid = ifelse(is_circle, value, NA)) %>%
  fill(circleid, .direction = "down") %>%
  filter(!is_circle) %>% # Keep only rows with node IDs
  rename(nodeid = value) %>%
  select(circleid, nodeid) # Reorder columns to match your target

# View the result
print(cir_final)

Output

# A tibble: 5 × 2
  circleid   nodeid
  <chr>      <chr> 
1 ASDDSDAS   111   
2 ASDDSDAS   132   
3 ASDDSDAS   456   
4 DSSADSDA   1     
5 DSSADSDA   12335 

This gives you the exact structure you need: each circle ID paired with its corresponding node IDs in two clearly named columns.

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