XML数据扁平化，如何找回原始数据

Question

Closed. This question needs to be more focused. It is not currently accepting answers.

---

想要改进这个问题吗？通过编辑这篇文章来更新问题，使其仅关注一个问题。

2 年前已关闭。

Answer 1

您提到您的现有字段位于另一个文件中。如果您可以使用它们来制作命名空白字符向量，如下所示：

existing_list <- c(id = "", v1 = "", v2 = "", v3 = "", v4 = "")

那么您可以执行以下操作：

df <- read.csv(text = "my_textfile.csv", header = FALSE)

id_list <- split(df, cumsum(df$V1 == "id"))

do.call(rbind, lapply(id_list, function(x) {
  vec <- setNames(x$V2, x$V1)
  existing_list[match(names(vec), names(existing_list))] <- vec
  as.data.frame(as.list(existing_list))
  }))

#>    id         v1         v2 v3         v4
#> 1 001 some_value                         
#> 2 002 some_value some_value              
#> 3 003            some_value              
#> 4 004                          some_value

再次强调，此处不使用外部包。

Answer 2

这是一种用 Python 实现的方法。

假设输入文件 'flattened_xml.txt' 包含：

id,001
v1,some_value1
id,002
v2,some_value2
v1,some_value3
id,003
v2,some_value4
id,004
v4,some_value5

代码：

# Get lines of text file.
with open('flattened_xml.txt') as file:
    data = file.read().splitlines()

# Determine what fiels are present.
fields = []
for line in data:
    id, _ = line.split(',')
    if id not in fields:
        fields.append(id)

table = [{field:field for field in fields}]  # Header row.
row = {}
for line in data:
    field, value = line.split(',')

    if field == fields[0]:
        row = {field: value}
        row.update({key: None for key in fields[1:]})
        continue

    row.update({key: value if key == field else None for key in fields[1:]})
    table.append(row)
    row = {key: None for key in fields}

for row in table:
    s = [f'{value if value is not None else "":11}' for value in row.values()]
    print('|' + '|'.join(s) + '|')

结果：

|id         |v1         |v2         |v4         |
|001        |some_value1|           |           |
|002        |           |some_value2|           |
|           |some_value3|           |           |
|003        |           |some_value4|           |
|004        |           |           |some_value5|