我就废话不多说了，直接上代码吧！

"""
对于每天存储文件，文件数量过多，占用空间
采用保存最新的三个文件
"""
from airflow import DAG
from airflow.operators.python_operator import PythonOperator
from airflow.models import Variable
from sctetl.airflow.utils import dateutils
from datetime import datetime,timedelta
import logging
import os
import shutil
"""
base_dir = "/data"
data_dir = "/gather"
"gather下边存在不同的文件夹"
"/data/gather/test"
"test路径下有以下文件夹"
"20180812、20180813、20180814、20180815、20180816"
"""

base_dir = Variable.get("base_dir")
data_dir = Variable.get("data_dir")
keep = 3

default_arg = {
  "owner":"airflow",
  "depends_on_past":False,
  "start_date":dateutils.get_start_date_local(2018,8,27,18,5),
  "email":[''],
  "email_on_failure":False,
  "email_on_retry":False,
  "retries":1,
  "retry_delay":timedelta(minutes=5)
}

dag = DAG(dag_id="keep_three_day",default_args=default_arg,schedule_interval=dateutils.get_schedule_interval_local(18,5))

def keep_three_day():
  path = os.path.join(base_dir, data_dir)
  date_cates = os.listdir(path)
  for cate in date_cates:
    p = os.path.join(base_dir, data_dir, cate)
    if os.path.isdir(p):
      dir_names = os.listdir(p)
      dir_names.sort()
      for i in dir_names[:-keep]:
        logging.info("删除目录 {path}".format(path=os.path.join(p, i)))
        shutil.rmtree(os.path.join(p, i))

with dag:
  keep_three_file = PythonOperator(task_id="keep_three_file",python_callable=keep_three_day(),dag=dag)

keep_three_file

以上这篇python 实现保存最新的三份文件,其余的都删掉就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多支持python博客。