收集精简的增量数据

用 Python 分析 IoT 数据

Matthias Voppichler

IT Developer

什么是缓存？

存储数据

数据流采集完成后
按观测逐条写入
- 会造成磁盘高负载
使用缓存

缓存

cache = []

def on_message(client, userdata, message):
    data = json.loads(message.payload)
    cache.append(data)

    if len(cache) > MAX_CACHE:

        with Path("data.txt").open("a") as f:
            f.writelines(cache)
        cache.clear()

# Connect function to mqtt datastream
subscribe.callback(on_message, 
                   topics="datacamp/energy", 
                   hostname=MQTT_HOST)

简易数据流

C331,6020
M640,104
C331,6129
M640,180
C331,6205
M640,256

观测时间戳

"payload 中的 timestamp"
message.timestamp
datetime.now()

观测时间戳

def on_message(client, userdata, message):
    publishtime = message.timestamp

    consume_time = datetime.utcnow()

pd.to_datetime()

print(df.head())

   timestamp         device  val
0  1540535443083       C331  347069.305500
1  1540535460858       C331  347069.381205

import pandas as pd
df["timestamp"] = pd.to_datetime(df["timestamp"], unit="ms")

  timestamp                  device  val
0 2018-10-26 06:30:43.083      C331  347069.305500
1 2018-10-26 06:31:00.858      C331  347069.381205

Passons à la pratique !

用 Python 分析 IoT 数据