尽可能减少 数据处理中的内存消耗

尽可能减少 数据处理中的内存消耗

服务器成本 时间成本

'''

{"ad_slots_id":1002,"uuid":"f18343c2-3e09-4abd-b3c5-e00cf33ff84d","industry_pid":0,"industry_id":0,"ip":"3661949473","site":72,"address":"https://info.b2b168.com/s168-54296673.html","create_date":"2019-01-02 14:56:58","ad_id":"33988392","uid":"33988392","keyword":"u71c3u70e7u673au914du4ef6","pageinfo":""}
{"ad_slots_id":1002,"uuid":"f18343c2-3e09-4abd-b3c5-e00cf33ff84d","industry_pid":0,"industry_id":0,"ip":"3661949473","site":72,"address":"https://info.b2b168.com/s168-54296673.html","create_date":"2019-01-02 14:56:58","ad_id":"50017820","uid":"50017820","keyword":"u5de5u4e1au6cb9u70dfu51c0u5316u5668","pageinfo":""}

'''

def fileRows(f, debug=False):
    l = []
    global pass_ip
    with open(f, 'r') as fr:
        for i in fr:
            try:
                # d = json.loads(i)
                i=i.strip('
')
                l.append(i)
            except Exception as e:
                if debug:
                    print(e)
                    print(i)
                    print(f)
    fr.close()
    del fr
    return l

    for f in file_list:
        if target_date not in f:
            continue
        rows_ = fileRows(f)
        print(f, ':', len(rows_))
        rows += rows_
        del rows_

    d = {}

    for i in rows:
        if 'uid' not in i:
            continue
        try:
            i = json.loads(i)
            uid, uuid, long_ip = i['uid'], i['uuid'], i['ip']
            if uid not in d:
                d[uid] = {}
                d[uid]['uuid'], d[uid]['long_ip'], d[uid]['pv'] = [], [], 0
            d[uid]['pv'] += 1
            d[uid]['uuid'].append(uuid)
            d[uid]['long_ip'].append(long_ip)
        except Exception as e:
            if 4 > 91:
                print(e)

　　

数据预处理阶段

数据的结构化处理会消耗不必要的内存，比如多行的json字符串构成的文件的逐行字符串转json

在数据的业务层面，逐行结构化，占用接近恒定的内存，增加对内存的控制性