如何拆分Pandas中多股票多层表头DataFrame为独立数据表

Hey there! No worries, splitting that multi-index DataFrame from stooq into separate DataFrames for each stock is totally straightforward. Here's exactly how you can get the output you're looking for:

Step-by-Step Solution #

First, let's recap your original data fetching code, then add the splitting logic using pandas' built-in xs() method (short for cross-section), which is perfect for working with multi-index columns.

import datetime as dt
import pandas as pd
import pandas_datareader.data as web

pd.set_option('display.max_columns', 500)
stocks = ["1.HK", "2.HK"]
start_time = dt.datetime(2022, 7, 1)
end_time = dt.datetime.now().date().isoformat()

# Fetch the multi-index data from stooq
df = web.DataReader(stocks, 'stooq', start_time, end_time)

# Split into individual DataFrames for each stock
# xs() extracts all columns for the specified symbol from the 'Symbols' level
df_1HK = df.xs('1.HK', axis=1, level='Symbols').sort_index()
df_2HK = df.xs('2.HK', axis=1, level='Symbols').sort_index()

# Print the results to verify
print("df_1HK:")
print(df_1HK)
print("\ndf_2HK:")
print(df_2HK)

What This Does #

• The xs('1.HK', axis=1, level='Symbols') call targets the second level of your column index (Symbols) and pulls all rows/columns associated with "1.HK"—this gives you exactly the Close, High, Low, Open, and Volume columns for that stock.
• sort_index() reorders the dates from oldest to newest, matching the format you showed in your desired output (since stooq returns data with the most recent date first by default).

Example Output #

Running this code will produce df_1HK and df_2HK structured exactly as you requested:

• Each uses Date as the index
• Columns are named Close, High, Low, Open, Volume
• All data is filtered to the respective stock

内容的提问来源于stack exchange，提问作者Marcin