使用Traffic库下载OpenSky航班数据触发MemoryError该如何解决？

DtypeWarning是pandas读取数据时自动推断列类型出错，对应列存在混合数据类型；MemoryError是单次拉取15天的全量数据量过大，超出内存分配上限，对应解决方案如下：

方案1：直接透传参数指定dtype #

opensky.history方法内部调用pandas做数据读取，所有pandas读取相关的参数都可以直接透传，你可以直接在调用时传入dtype和low_memory参数即可：

from traffic.data import opensky, airports

# 自定义dtype映射，把非必要的float64转成float32节省内存，同时指定混合类型列的类型
# 你可以先拉取1小时的小样本确认第18列的字段名，再对应设置类型
dtype_map = {
    "onground": bool,
    "latitude": "float32",
    "longitude": "float32",
    "altitude": "float32",
    "velocity": "float32",
    "heading": "float32",
    "vertrate": "float32",
    "callsign": str,
    "icao24": str
}

airport = airports['LEMD']
flight = opensky.history(
    start="2021-06-01 00:00",
    stop="2021-06-16 00:00",
    arrival_airport=airport.icao,
    low_memory=False,
    dtype=dtype_map
)
flight.to_csv(r"C:\Users\peter\Desktop\cosicas.csv")

方案2：分批次拉取（更稳定，推荐） #

如果你的内存还是不足以支撑全量数据，可以按天拆分拉取，每次拉取一天的数据追加到csv文件，拉完立即释放内存，避免一次性占用过高内存：

from traffic.data import opensky, airports
from datetime import datetime, timedelta

airport = airports['LEMD']
start_date = datetime(2021,6,1)
end_date = datetime(2021,6,16)
csv_path = r"C:\Users\peter\Desktop\cosicas.csv"
dtype_map = {
    "onground": bool,
    "latitude": "float32",
    "longitude": "float32",
    "altitude": "float32",
    "velocity": "float32",
    "heading": "float32",
    "vertrate": "float32",
    "callsign": str,
    "icao24": str
}

current_date = start_date
first_write = True
while current_date < end_date:
    next_date = current_date + timedelta(days=1)
    daily_flight = opensky.history(
        start=current_date.strftime("%Y-%m-%d %H:%M"),
        stop=next_date.strftime("%Y-%m-%d %H:%M"),
        arrival_airport=airport.icao,
        low_memory=False,
        dtype=dtype_map
    )
    # 追加写入csv
    daily_flight.to_csv(
        csv_path,
        mode='w' if first_write else 'a',
        header=first_write,
        index=False
    )
    first_write = False
    current_date = next_date
    # 手动释放内存
    del daily_flight

内容的提问来源于stack exchange，提问作者Peter La Anguila