如何用Python绘制大型CSV文件数据？Quandl股票可视化求助

First off, let's break down the issues in your code that are causing it to fail, then build a working version that includes the features you need: selecting Date/Close Price columns and filtering the last 100 trading days.

Key Issues in Your Original Code #

• Syntax Errors:
  • With should be lowercase with (Python is case-sensitive)
  • You used Chinese quotation marks ‘rb’ instead of English ones 'rb'
  • print row is Python 2 syntax; use print(row) for Python 3
  • You named a variable csv, which overwrites the built-in csv module—this breaks all subsequent CSV operations
• Logical Missteps:
  • You're repeatedly fetching the same Quandl URL multiple times (inefficient and unnecessary)
  • np.loadtxt can't directly parse remote URLs with complex date formats correctly, and you set the wrong delimiter (Quandl uses commas, not spaces)
  • You tried to mix csv.reader and csv.writer on the same file handle, which will cause conflicts
  • No logic to filter the last 100 trading days

Fixed Version (Using Base Python Libraries) #

This version fixes all bugs, fetches the data once, extracts Date/Close columns, filters the last 100 rows, and plots the data:

import urllib.request
import csv
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import style
from datetime import datetime

style.use('fivethirtyeight')

# Fetch data once and process directly (no need to save to file unless you want to)
url = 'https://www.quandl.com/api/v3/datasets/EOD/V.csv?api_key=Fa1P1yZLGnGSsXktrvzL'
response = urllib.request.urlopen(url)
csv_data = response.read().decode('utf-8').splitlines()

# Parse CSV, skip header, extract Date and Close columns
reader = csv.reader(csv_data)
next(reader)  # Skip the header row
rows = list(reader)

# Filter last 100 trading days
last_100_rows = rows[-100:]

# Convert data to plot-ready formats
dates = []
close_prices = []
for row in last_100_rows:
    # Parse Quandl's date format (YYYY-MM-DD)
    date_obj = datetime.strptime(row[0], '%Y-%m-%d')
    dates.append(date_obj)
    close_prices.append(float(row[4]))  # Close price is the 5th column (0-based index 4)

# Plot the data
plt.figure(figsize=(10,6))
plt.plot(dates, close_prices, label='Visa Close Price')
plt.title('Visa EOD Stock Prices (Last 100 Trading Days)')
plt.ylabel('Stock Price (USD)')
plt.xlabel('Date')
plt.xticks(rotation=45)
plt.legend()
plt.tight_layout()  # Fixes label cutoff issues
plt.show()

Optimized Version (Using Pandas) #

For time-series stock data, Pandas is far more efficient and readable. It handles date parsing, column selection, and filtering in just a few lines:

import pandas as pd
import matplotlib.pyplot as plt
from matplotlib import style

style.use('fivethirtyeight')

# Fetch and parse data directly with Pandas
url = 'https://www.quandl.com/api/v3/datasets/EOD/V.csv?api_key=Fa1P1yZLGnGSsXktrvzL'
df = pd.read_csv(url, parse_dates=['Date'], index_col='Date')

# Select Close column and filter last 100 rows
last_100_days = df['Close'].tail(100)

# Plot
plt.figure(figsize=(10,6))
last_100_days.plot(label='Visa Close Price')
plt.title('Visa EOD Stock Prices (Last 100 Trading Days)')
plt.ylabel('Stock Price (USD)')
plt.xlabel('Date')
plt.legend()
plt.tight_layout()
plt.show()

Quick Notes #

• In the Quandl EOD dataset, the Close column is at index 4 (0-based) in the raw CSV
• Pandas automatically parses dates and handles time-series indexing, which simplifies filtering with tail(100)
• Both versions avoid overwriting built-in module names (like csv) to prevent conflicts

内容的提问来源于stack exchange，提问作者Ross Irving