代码1

import json
import time
import pandas as pd

data_sets = pd.read_pickle("val_token_id.pandas_pickle")
data_sets=[str(i) for i in data_sets]
start=time.time()
[json.loads(i) for i in data_sets]
print(time.time()-start)

start=time.time()
[eval(i) for i in data_sets]
print(time.time()-start)

结果图

代码2

import json
import time
from multiprocessing import Process, Manager, freeze_support
import pandas as pd
from tqdm import tqdm
def json_loads_data(return_list,one_data):
    return_list+=[json.loads(i)  for i in tqdm(one_data)]

if __name__ == '__main__':
    freeze_support()
    data_sets = pd.read_pickle("val_token_id.pandas_pickle")
    data_sets = [str(i) for i in data_sets]

    start = time.time()
    data = Manager().list()
    num = 5
    p_list = []
    for i in range(0, len(data_sets), len(data_sets)//num):
        j = i + len(data_sets)//num
        p = Process(target=json_loads_data, args=(data, data_sets[i:j]))
        p.start()
        p_list.append(p)
    for p in p_list:
        p.join()

    print("multi_json_loads", time.time() - start)

    start = time.time()
    [json.loads(i) for i in data_sets]
    print("json_loads", time.time() - start)

    start = time.time()
    pd.DataFrame(data_sets)[0].apply(lambda x: json.loads(x)).values.tolist()
    print("dataFrame_apply", time.time() - start)

    start = time.time()
    json.loads(str(data_sets).replace("'", ""))
    print("json_loads_str", time.time() - start)

    start = time.time()
    [eval(i) for i in data_sets]
    print("eval", time.time() - start)

参考地址

https://blog.csdn.net/qq_35869630/article/details/105919104
Python 在大数据处理下的优化（一）用json.loads比eval快10倍！