如何根据 pandas 中其他列的条件创建一个新列? [英] How do I create a new column based on conditions of other columns in pandas?
问题描述
我有一个熊猫数据框,我想在其中创建一个新列,其值基于其他三个列的值.
I have a pandas dataframe in which I want to create a new columns, which values are based on values of three other columns.
首先创建一个列,并给它一个随机值300:
first a created the column and just gave it a random value 300:
data['stability'] = 300
然后我设置条件:
data['stability'][(data['wind_speed'] <= 3) & (data['clouds'] < 4 ) & (data['dagnacht'] == 'nacht')] = 6
data['stability'][(data['wind_speed'] > 3) & (data['clouds'] < 4 ) & (data['dagnacht'] == 'nacht')] = 5
data['stability'][(data['wind_speed'] >= 5) & (data['clouds'] < 4 ) & (data['dagnacht'] == 'nacht')] = 4
data['stability'][(data['wind_speed'] <= 3) & (data['clouds'] >= 4 ) & (data['dagnacht'] == 'nacht')] = 5
data['stability'][(data['wind_speed'] > 3) & (data['clouds'] >= 4 ) & (data['dagnacht'] == 'nacht')] = 4
如果您检查条件是否存在,则它确实表明条件存在: 输入:
If you check if the condition exist, it does tell that the conditions exist: input:
data['stability'][(data['wind_speed'] > 3) & (data['clouds'] >= 4 ) & (data['dagnacht'] == 'nacht')]
输出:
2011-08-04 21:00:00 300.0
2011-08-04 22:00:00 300.0
2011-08-04 23:00:00 300.0
2011-08-05 00:00:00 300.0
2011-08-05 01:00:00 300.0
2011-08-05 02:00:00 300.0
2011-08-05 03:00:00 300.0
2011-08-05 04:00:00 300.0
2011-08-05 05:00:00 300.0
2011-08-06 23:00:00 300.0
2011-08-07 00:00:00 300.0
2011-08-07 01:00:00 300.0
但是正如您所看到的,它仍然具有我一开始就给它的值300,而不是我现在想要的值4. value_counts给我300作为唯一值 由于某种原因,它可以读取条件,但不会将新值分配给稳定性.
But as you can see it still had the value of 300 I gave it in the beginning, and not the value 4 I want to have it now. value_counts gives me 300 as the only value For some reason it can read to condition but does not assign the new value to stability.
我正在使用python 2.7和pandas 0.18.0
I am working with python 2.7 and pandas 0.18.0
我的数据集如下:
wind_speed clouds stability dagnacht
date
2016-03-21 19:00:00 4.73 7 300.0 nacht
2016-03-21 19:10:00 4.58 NaN 300.0 nacht
2016-03-21 19:20:00 4.75 NaN 300.0 nacht
2016-03-21 19:30:00 3.67 NaN 300.0 nacht
2016-03-21 19:40:00 3.41 NaN 300.0 nacht
2016-03-21 19:50:00 3.61 NaN 300.0 nacht
2016-03-21 20:00:00 3.31 8 300.0 nacht
2016-03-21 20:10:00 3.30 NaN 300.0 nacht
2016-03-21 20:20:00 3.39 NaN 300.0 nacht
2016-03-21 20:30:00 3.59 NaN 300.0 nacht
2016-03-21 20:40:00 3.24 NaN 300.0 nacht
2016-03-21 20:50:00 2.99 NaN 300.0 nacht
2016-03-21 21:00:00 3.04 7 300.0 nacht
2016-03-21 21:10:00 3.01 NaN 300.0 nacht
2016-03-21 21:20:00 2.63 NaN 300.0 nacht
2016-03-21 21:30:00 2.41 NaN 300.0 nacht
2016-03-21 21:40:00 2.42 NaN 300.0 nacht
2016-03-21 21:50:00 2.49 NaN 300.0 nacht
2016-03-21 22:00:00 2.31 8 300.0 nacht
2016-03-21 22:10:00 2.24 NaN 300.0 nacht
2016-03-21 22:20:00 1.89 NaN 300.0 nacht
2016-03-21 22:30:00 1.88 NaN 300.0 nacht
2016-03-21 22:40:00 1.83 NaN 300.0 nacht
2016-03-21 22:50:00 1.83 NaN 300.0 nacht
2016-03-21 23:00:00 1.86 8 300.0 nacht
2016-03-21 23:10:00 2.29 NaN 300.0 nacht
2016-03-21 23:20:00 2.53 NaN 300.0 nacht
2016-03-21 23:30:00 2.36 NaN 300.0 nacht
2016-03-21 23:40:00 2.04 NaN 300.0 nacht
2016-03-21 23:50:00 1.83 NaN 300.0 nacht
预先感谢您的帮助
推荐答案
You're performing chained indexing, change your lines to this form:
data.loc[(data['wind_speed'] > 3) & (data['clouds'] >= 4 ) & (data['dagnacht'] == 'nacht'), 'stability'] = 4
因此,您可以根据数据视图而不是副本视图进行操作
So you operate on a view on your data rather than a copy
In [19]:
data.loc[(data['wind_speed'] > 3) & (data['clouds'] >= 4 ) & (data['dagnacht'] == 'nacht'), 'stability'] = 4
data
Out[19]:
wind_speed clouds stability dagnacht
date
2016-03-21 19:00:00 4.73 7.0 4.0 nacht
2016-03-21 19:10:00 4.58 NaN 300.0 nacht
2016-03-21 19:20:00 4.75 NaN 300.0 nacht
2016-03-21 19:30:00 3.67 NaN 300.0 nacht
2016-03-21 19:40:00 3.41 NaN 300.0 nacht
2016-03-21 19:50:00 3.61 NaN 300.0 nacht
2016-03-21 20:00:00 3.31 8.0 4.0 nacht
2016-03-21 20:10:00 3.30 NaN 300.0 nacht
2016-03-21 20:20:00 3.39 NaN 300.0 nacht
2016-03-21 20:30:00 3.59 NaN 300.0 nacht
2016-03-21 20:40:00 3.24 NaN 300.0 nacht
2016-03-21 20:50:00 2.99 NaN 300.0 nacht
2016-03-21 21:00:00 3.04 7.0 4.0 nacht
2016-03-21 21:10:00 3.01 NaN 300.0 nacht
2016-03-21 21:20:00 2.63 NaN 300.0 nacht
2016-03-21 21:30:00 2.41 NaN 300.0 nacht
2016-03-21 21:40:00 2.42 NaN 300.0 nacht
2016-03-21 21:50:00 2.49 NaN 300.0 nacht
2016-03-21 22:00:00 2.31 8.0 300.0 nacht
2016-03-21 22:10:00 2.24 NaN 300.0 nacht
2016-03-21 22:20:00 1.89 NaN 300.0 nacht
2016-03-21 22:30:00 1.88 NaN 300.0 nacht
2016-03-21 22:40:00 1.83 NaN 300.0 nacht
2016-03-21 22:50:00 1.83 NaN 300.0 nacht
2016-03-21 23:00:00 1.86 8.0 300.0 nacht
2016-03-21 23:10:00 2.29 NaN 300.0 nacht
2016-03-21 23:20:00 2.53 NaN 300.0 nacht
2016-03-21 23:30:00 2.36 NaN 300.0 nacht
2016-03-21 23:40:00 2.04 NaN 300.0 nacht
2016-03-21 23:50:00 1.83 NaN 300.0 nacht
如果您尝试使用代码,则应该提出警告:
A warning should've been raised if you tried your code:
In [21]:
data['stability'][(data['wind_speed'] > 3) & (data['clouds'] >= 4 ) & (data['dagnacht'] == 'nacht')] = 4
data
C:\WinPython-64bit-3.4.3.5\python-3.4.3.amd64\lib\site-packages\IPython\kernel\__main__.py:1: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
if __name__ == '__main__':
这篇关于如何根据 pandas 中其他列的条件创建一个新列?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!