使用urllib与Pandas抓取HTML表格转CSV时遇TypeError错误求助

Hey there, let's break down why you're running into this TypeError and fix it up quickly.

你的问题场景 #

你尝试抓取Yahoo Finance的BTC-JPY历史数据表格并保存为CSV，但代码报错：

代码如下：

import urllib, pandas as pd
url = 'https://finance.yahoo.com/quote/BTC-JPY/history?period1=1314403200&period2=1314489600&interval=1d&filter=history&frequency=1d'
fo = 'test.txt'
response = urllib.request.urlopen(url)
html = response.read()
data = pd.read_html(html)
data.to_csv(fo, index = False, header=False, sep=',', mode='w')

错误信息：TypeError: ufunc 'add' did not contain a loop with signature matching...

问题根源 #

The core issue here is that pd.read_html() doesn't return a single DataFrame—it returns a list of DataFrames (since an HTML page can have multiple tables). When you try to call .to_csv() directly on this list, pandas can't process it, hence the confusing type error.

修正后的代码 #

Here's the working version with explanations:

import urllib.request
import pandas as pd

# Clean up the URL (replace & with actual &)
url = 'https://finance.yahoo.com/quote/BTC-JPY/history?period1=1314403200&period2=1314489600&interval=1d&filter=history&frequency=1d'
# Use .csv extension for proper CSV file
output_file = 'btc_jpy_history.csv'

# Fetch the HTML content
response = urllib.request.urlopen(url)
html = response.read()

# Extract tables from HTML (returns a list)
tables = pd.read_html(html)

# Check if we found any tables, then use the first one (Yahoo's history table is first)
if tables:
    target_df = tables[0]
    # Save to CSV - keep headers since they're useful for the data
    target_df.to_csv(output_file, index=False, header=True, sep=',', mode='w')
    print(f"Data saved successfully to {output_file}!")
else:
    print("No tables found in the fetched HTML.")

额外提示 #

• I switched the output file to .csv instead of .txt—this makes the file recognizable as a CSV for other tools.
• Added a check for empty tables to avoid another error if Yahoo's page structure changes.
• Fixed the URL entities (& → &) for cleaner, more reliable parsing.

内容的提问来源于stack exchange，提问作者Roman