如何将带有空白换行符的 CSV 数据块解析为数据帧列表答案

【问题标题】：How to parse chunk of CSV data with white space line breaks into list of data frames如何将带有空白换行符的 CSV 数据块解析为数据帧列表
【发布时间】：2021-12-27 06:44:20
【问题描述】：

我有以下文本文件，它包含几块表。每个块由空格分隔。

GENERALIZED BORN:
Complex Energy Terms
Frame #,BOND,ANGLE
0,6603.0521,7264
1,7434.9885,7602

Receptor Energy Terms
Frame #,BOND,ANGLE
0,6140.6338,5383.1241
1,6885.2965,5653.6637

Ligand Energy Terms
Frame #,BOND,ANGLE
0,462.4183,1881.428
1,549.692,1949.0482

如何使用 R 将此单个文本文件解析为三个数据框或小标题的列表？

我试过了，但失败了：

library(readr)
readr::read_lines_chunked("myfile.txt", skip =1, chunk_size = 4)

因为readr::read_lines_chunked 无法识别块之间的空白分隔符。

【问题讨论】：

标签： r tidyverse readr

【解决方案1】：

我不清楚该功能是否应该像那样使用。我的猜测不是。但是您可以手动解析数据。创建一个列表，解析出块并将其保存到一个列表中。

xy <- readLines(con = "test.txt")
xy <- xy[-1]  # remove GENERALIZED BORN

xy <- xy[which(xy != "")]
# Start of a break is needed for names and subsetting in a loop.
breaks <- which(grepl("^.*Energy Terms", x = xy))

dfs <- vector(mode = "list", length = length(breaks))
names(dfs) <- xy[breaks]

# Adding one accounts for the Energy Terms line. It's either here
# or in the loop.
chunks <- breaks + 1

for (chunk in seq_along(chunks)) {
  # If we extract the name and use it to subset, the order of the
  # dfs doesn't really matter.
  chk.name <- xy[chunks[chunk] - 1]
  
  from <- chunks[chunk]
  to <- chunks[chunk + 1] - 2
  
  # When working with the last chunk, this sets the end of the text.
  if (is.na(to)) {
    to <- length(xy)
  }
  
  chk <- xy[from:to]

  tmp <- read.table(
    text =  paste(chk, collapse = "\n"), 
    header = TRUE, 
    comment.char = "", 
    sep = ","
    )
  
  dfs[[chk.name]] <- tmp
}

结果

$`Complex Energy Terms`
  Frame..     BOND ANGLE
1       0 6603.052  7264
2       1 7434.989  7602

$`Receptor Energy Terms`
  Frame..     BOND    ANGLE
1       0 6140.634 5383.124
2       1 6885.297 5653.664

$`Ligand Energy Terms`
  Frame..     BOND    ANGLE
1       0 462.4183 1881.428
2       1 549.6920 1949.048

【讨论】：