Question

numeric_cols = ['temp', 'windchill', 'dewpoint', 'humidity', 'pressure', 'visibility', 'wind_speed', 'gust_speed', 'precip']
weather_list[numeric_cols] = weather_list[numeric_cols].apply(lambda x: re.sub('[^0-9]', '', str(x)))
weather_list[numeric_cols] = pd.to_numeric(weather_list[numeric_cols], errors='coerce')
weather_list[numeric_cols] = weather_list[numeric_cols] / 10

我试图对数据集进行一些清理但是我遇到了形状不匹配错误。由于错误，它说我匹配30列，0行，9列，30行......我明显做错了！我以前曾经使用过这种方法几次但从来没有错过 - 任何人都对我做错了什么有任何建议？数据从html拉到pandas df＆＃39; weather_list＆＃39;。

time     temp windchill dewpoint humidity  pressure visibility  \
0   12:53 AM  21.0 °F         -  19.0 °F      92%  30.47 in    10.0 mi   
1    1:53 AM  21.9 °F         -  19.9 °F      92%  30.48 in    10.0 mi   
2    2:53 AM  21.9 °F         -  19.0 °F      89%  30.50 in    10.0 mi   
3    3:53 AM  21.0 °F         -  19.0 °F      92%  30.50 in    10.0 mi   
4    4:53 AM  19.9 °F         -  18.0 °F      92%  30.51 in    10.0 mi   
5    5:53 AM  21.0 °F         -  18.0 °F      88%  30.51 in    10.0 mi   


   wind_direction wind_speed gust_speed  precip  events conditions  
0            Calm       Calm          -     NaN     NaN      Clear  
1            Calm       Calm          -     NaN     NaN      Clear  
2            Calm       Calm          -     NaN     NaN      Clear  
3            Calm       Calm          -     NaN     NaN      Clear  
4            Calm       Calm          -     NaN     NaN      Clear  
5            Calm       Calm          -     NaN     NaN      Clear

谢谢！

Answer 1

我们可以使用applymap将函数应用于多个列。请对您的代码进行以下更改。

## using applymap here
weather_list[numeric_cols] = weather_list[numeric_cols].applymap(lambda x: re.sub(r'[^0-9]', '', str(x)))

## now we pass series to pd.to_numeric instead of data frame
weather_list[numeric_cols] = weather_list[numeric_cols].apply(lambda x: pd.to_numeric(x, errors='coerce'))

weather_list[numeric_cols] = weather_list[numeric_cols] / 10

ValueError：形状不匹配：形状的值数组（30，）无法广播到形状的索引结果（9,30）

1 个答案: