pandas 分离pandas DataFrame的正负值
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/35204716/
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
Separating positive and negative values of pandas DataFrame
提问by David Hancock
I need to separately sum all positive and negative values in a column ie
我需要分别对列中的所有正值和负值求和，即
pos_values = [x for x in df.prediction_at_ohlcv_end_date if x > 0]
neg_values = [x for x in df.prediction_at_ohlcv_end_date if x < 0]
Here's a data sample
这是一个数据样本
market_trading_pair next_future_timestep_return ohlcv_start_date prediction_at_ohlcv_end_date
0 Poloniex_ETH_BTC 0.003013 1450753200 -0.157053
1 Poloniex_ETH_BTC -0.006521 1450756800 -0.920074
2 Poloniex_ETH_BTC 0.003171 1450760400 0.999806
3 Poloniex_ETH_BTC -0.003083 1450764000 0.627140
4 Poloniex_ETH_BTC -0.001382 1450767600 0.999857
What's a nice way to do this in pandas ?
在Pandas中这样做的好方法是什么？
EDIT:
编辑：
I have been able to do this thanks to some helpful stackers, I realised I can't a futher calculation however. `
多亏了一些有用的堆垛机，我才能够做到这一点，但我意识到我无法进一步计算。`
if prediction_at_ohlcv_end_date > 0 :
return = prediction_at_ohlcv_end_date * next_future_timestep_return.
For each element in the frame, Any ideas?`
对于框架中的每个元素，有什么想法吗？`
回答by Anton Protopopov
You could use method sumof pandas.Seriesfor your particular column:
你可以使用方法sum的pandas.Series为您的特定列：
neg = df.prediction_at_ohlcv_end_date[df.prediction_at_ohlcv_end_date < 0].sum()
pos = df.prediction_at_ohlcv_end_date[df.prediction_at_ohlcv_end_date >= 0].sum()
In [51]: pos
Out[51]: 2.6268029999999998
In [52]: neg
Out[52]: -1.077127
For your values:
对于您的价值观：
pos_values = df.prediction_at_ohlcv_end_date[df.prediction_at_ohlcv_end_date >= 0]
neg_values = df.prediction_at_ohlcv_end_date[df.prediction_at_ohlcv_end_date < 0]
EDIT
编辑
For your edit you could do:
对于您的编辑，您可以执行以下操作：
mask = df.prediction_at_ohlcv_end_date >= 0
res = df.prediction_at_ohlcv_end_date[mask] * df.next_future_timestep_return[mask]
In [10]: res
Out[10]:
2 0.003170
3 -0.001933
4 -0.001382
dtype: float64
