代码之家 › 专栏 › 技术社区 › Hariom Singh

使用panda解析XML

pandas xml python

Hariom Singh · 技术社区 · 7 年前

尝试解析XML,然后将其表示为熊猫数据帧

<?xml version="1.0"?><results>
<header>
  <cloc_url>github.com/AlDanial/cloc</cloc_url>
  <cloc_version>1.74</cloc_version>
  <elapsed_seconds>0.940369129180908</elapsed_seconds>
  <n_files>124</n_files>
  <n_lines>8440</n_lines>
  <files_per_second>131.863112209998</files_per_second>
  <lines_per_second>8975.19892784178</lines_per_second>
  <report_file>/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem_cloc.xml</report_file>
</header>
<files>
  <file name="/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem/aem-parent/pom.xml" blank="13" comment="23" code="491"  language="Maven" />
  <file name="/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem/aem-core/aem-core-bundle/src/test/resources/assets.json" blank="0" comment="0" code="357"  language="JSON" />
  <file name="/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem/aem-core/aem-core-bundle/src/main/java/com/chute/aem/core/api/impl/UserServiceImpl.java" blank="26" comment="21" code="202"  language="Java" />

输出类似

file name                                 blank  comment language code
Repo/ignite-chute-aem/aem-parent/pom.xml"  "13"   "23"     Maven   491
<fullpath>/assets.json"                     "12"   "3"      c       432

我只会写几行

import pandas as pd
from xml.etree import ElementTree
tree = ElementTree.parse('/Users/hariomsingh/Desktop/individualxml/ignite-chute-aem_cloc.xml')
root = tree.getroot()

print(root)
print(tree.iter())

csv_data = []
fields =  ['file name','blank','comment', 'language', 'code']

1 回复 | 直到 7 年前

TSeymour 7 年前

假设你对安装漂亮的soup4(即, pip3 install beautifulsoup4 )以及熊猫(即, pip3 install pandas ,那么这应该可以做到:

from bs4 import BeautifulSoup as Soup
import pandas

xml = """
<?xml version="1.0"?><results>
<header>
  <cloc_url>github.com/AlDanial/cloc</cloc_url>
  <cloc_version>1.74</cloc_version>
  <elapsed_seconds>0.940369129180908</elapsed_seconds>
  <n_files>124</n_files>
  <n_lines>8440</n_lines>
  <files_per_second>131.863112209998</files_per_second>
  <lines_per_second>8975.19892784178</lines_per_second>
  <report_file>/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem_cloc.xml</report_file>
</header>
<files>
  <file name="/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem/aem-parent/pom.xml" blank="13" comment="23" code="491"  language="Maven" />
  <file name="/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem/aem-core/aem-core-bundle/src/test/resources/assets.json" blank="0" comment="0" code="357"  language="JSON" />
  <file name="/Users/hariomsingh/Desktop/ignitechute/Repo/ignite-chute-aem/aem-core/aem-core-bundle/src/main/java/com/chute/aem/core/api/impl/UserServiceImpl.java" blank="26" comment="21" code="202"  language="Java" />
"""

soup = Soup(xml, 'lxml')

records = []

for file in soup.findAll('file'):
    records.append(file.attrs)

data_table = pandas.DataFrame(records)

# this prints the table without the long file name to ease seeing all other fields
print(data_table.drop('name', axis=1))

# this prints just the names (or at least the bit that pandas prints by default)
print(data_table['name'])

# saving them to disk so you can see the entire table in excel or similar
data_table.to_csv('output.csv', index=False)