如何从API正确解析和处理Apache Parquet数据

import requests
import io
import pyarrow.parquet as pq
import pandas as pd # 假设后续会用到Pandas

def get_orders_data_incorrect(date):
    url = "YOUR_API_BASE_URL/orders" # 替换为你的API地址
    headers = {}
    params = {
        "date": date
    }

    response = requests.get(url, headers=headers, params=params)

    if response.status_code == 200:
        # 错误的做法：将二进制数据强制解码为文本字符串
        data_str = response.text.strip()
        return data_str
    else:
        print(f"Failed to fetch orders data: {response.status_code}")
        return None

# 调用函数并尝试解码
date_to_fetch = "2023-12-08"
orders_info_str = get_orders_data_incorrect(date_to_fetch)

if orders_info_str:
    try:
        # 尝试将损坏的字符串编码回字节流，然后解析
        buffer = io.BytesIO(orders_info_str.encode())
        table = pq.read_table(buffer) # 这一行通常会报错
        df = table.to_pandas()
        print(df.head())
    except Exception as e:
        print(f"Error decoding Parquet: {e}")
        # 错误信息可能类似于 'Parquet magic bytes not found' 或其他与格式相关的错误

import requests
import io
import pandas as pd

def get_orders_data_pandas(date: str) -> pd.DataFrame | None:
    """
    从API获取Parquet数据并使用Pandas直接解析。
    """
    url = "YOUR_API_BASE_URL/orders" # 替换为你的API地址
    headers = {} # 根据需要添加认证或其他头信息
    params = {
        "date": date
    }

    try:
        response = requests.get(url, headers=headers, params=params, stream=False) # stream=False确保完整下载
        response.raise_for_status() # 检查HTTP请求是否成功 (2xx状态码)

        # 核心：直接使用 response.content 获取原始字节流
        # 并通过 io.BytesIO 封装成文件对象供 pandas.read_parquet 读取
        df = pd.read_parquet(io.BytesIO(response.content))
        return df
    except requests.exceptions.RequestException as e:
        print(f"API请求失败: {e}")
        return None
    except Exception as e:
        print(f"解析Parquet数据失败: {e}")
        return None

# 示例调用
date_to_fetch = "2023-12-08"
orders_df = get_orders_data_pandas(date_to_fetch)

if orders_df is not None:
    print("成功获取并解析订单数据，前5行：")
    print(orders_df.head())
    print(f"DataFrame的形状: {orders_df.shape}")
else:
    print("未能获取或解析订单数据。")

import requests
import io
import pyarrow.parquet as pq
import pandas as pd

def get_orders_data_pyarrow(date: str) -> pd.DataFrame | None:
    """
    从API获取Parquet数据并使用PyArrow解析。
    """
    url = "YOUR_API_BASE_URL/orders" # 替换为你的API地址
    headers = {} # 根据需要添加认证或其他头信息
    params = {
        "date": date
    }

    try:
        response = requests.get(url, headers=headers, params=params, stream=False)
        response.raise_for_status()

        # 核心：使用 response.content 获取原始字节流
        buffer = io.BytesIO(response.content)

        # 使用 pyarrow.parquet.read_table 从字节流中读取Parquet表
        table = pq.read_table(buffer)

        # 将PyArrow表转换为Pandas DataFrame
        df = table.to_pandas()
        return df
    except requests.exceptions.RequestException as e:
        print(f"API请求失败: {e}")
        return None
    except Exception as e:
        print(f"解析Parquet数据失败: {e}")
        return None

# 示例调用
date_to_fetch = "2023-12-08"
orders_df_pyarrow = get_orders_data_pyarrow(date_to_fetch)

if orders_df_pyarrow is not None:
    print("成功获取并解析订单数据 (通过PyArrow)，前5行：")
    print(orders_df_pyarrow.head())
    print(f"DataFrame的形状: {orders_df_pyarrow.shape}")
else:
    print("未能获取或解析订单数据。")