python - “模块”对象没有属性 'isna'

Question

我有一个名为 batch_df 的 df，我在 df 中添加了“as_percentage”。如果 hours 是 Nan，则将 as_percentage 设置为 0。代码在 PyCharm 中运行良好，但是当我在终端上运行时，出现此错误:请帮忙!

Traceback (most recent call last):

File "/Users/vic_ong/dev/resource-management/generateBatch.py", line 166, in main()

File "/Users/vic_ong/dev/resource-management/generateBatch.py", line 51, in main

batch_df.loc[pd.isna(batch_df['hours']) == False, 'as_percentage'] = 0 AttributeError: 'module' object has no attribute 'isna'

编辑——我的代码如下所示:

#!/usr/bin/env python2
import lib
import pandas as pd
import datetime
import os
pd.set_option('expand_frame_repr', False)
# get directories
batchInputDir = '/Users/vic_ong/dev/resource-management/data/input/batchUpload'


def main():
# start timer
start_time = datetime.datetime.now()

# walk into every csv file in batchUpload and concatenate them into one df
list_ = []
for root, dirs, files in os.walk(batchInputDir):
    for file in files:
        if file.endswith('.csv'):
            df = pd.read_csv(os.path.join(batchInputDir, file), index_col=None, header=0)
            list_.append(df)
batch_df = pd.concat(list_)
batch_df = batch_df.dropna(how='all')
batch_df.reset_index(drop=True, inplace=True)
# clean data frame
# rename to standardized column names
batch_df.rename(columns={'Resource Name': 'name',
                         'Hours': 'hours'},
                inplace=True)
# determine if value given was a percentage
batch_df['as_percentage'] = 1
batch_df.loc[pd.isna(batch_df['hours']) == False, 'as_percentage'] = 0
print batch_df.head

if __name__ == '__main__':
main()

batch_df 样本:

将小时命名为_percentage
维南 1
vic1 南 1

Answer 1

The isna alias for isnull was only added in version 0.21 (0.22 是截至 2018 年 1 月的最新版本)；在 0.20 及更早版本中，它不存在；您必须改用旧的(且仍受支持)isnull 名称。更新您的 pandas 安装，或使用旧名称。