Python Pandas：仅从某些列创建新数据框
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/36518027/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Pandas: Creating new data frame from only certain columns
提问by ValientProcess
I have a csv file with measurements, and I want to create a new csv file with the hourly averages and standard deviations. But only for certain columns.
我有一个带有测量值的 csv 文件，我想创建一个包含每小时平均值和标准偏差的新 csv 文件。但仅适用于某些列。
Example:
例子：
csv1:
csv1：
YY-MO-DD HH-MI-SS_SSS | Acceleration | Lumx | Pressure
2015-12-07 20:51:06:608 | 22.7 | 32.3 | 10
2015-12-07 20:51:07:609 | 22.5 | 47.7 | 15
to csv 2 (only for the pressure and acceleration:
到 csv 2（仅用于压力和加速度：
YY-MO-DD HH-MI-SS_SSS | Acceleration avg | Pressure avg
2015-12-07 20:00:00:000 | 22.6 | 12.5
2015-12-07 21:00:00:000 | .... | ....
Now I have an idea (thanks to the people on this site) on how to calculate the averages - but i'm having trouble on creating a new smaller dataframe that contains the calculations for a few columns.
现在我有一个关于如何计算平均值的想法（感谢本网站上的人） - 但我在创建一个包含几列计算的新的较小数据框时遇到了麻烦。
Thanks !!!
谢谢 ！！！
回答by su79eu7k
You should make smaller df like below,
你应该像下面那样制作更小的 df，
csv2 = csv1[['Acceleration', 'Pressure']].copy()
and can handle the csv2. (You said you have an idea about avg calculation)
FYI, .copy()could be omitted if you are sure about view versus copy.
并且可以处理csv2。（你说你有一个关于 avg 计算的想法）仅供参考，.copy()如果你确定view 与 copy，可以省略。
回答by leerssej
csv2 = csv1.loc[:, ['Acceleration', 'Pressure']]
.loc[]helps keep the subsetting operation explicit and consistent..loc[]always returns a copy so the original dataframe is never modified.
.loc[]有助于保持子集操作的明确和一致。.loc[]总是返回一个副本，所以原始数据帧永远不会被修改。
(for further discussion and great examples of the different view vs. copyalternatives please see: Pandas: Knowing when an operation affects the original dataframe)
（有关不同view vs. copy替代方案的进一步讨论和很好的示例，请参阅：Pandas：了解操作何时影响原始数据帧）
