如何在条件保持为真时循环保持值不变

Question

规则是：

如果Regime = 0 => Position = 0
如果Regime = 1 => Position = Size
如果Regime = -1=> Position = -Size

一旦Position 取一个值，我想保持该值不变，而Re 保持为1 或-1

任何想法如何保持Position 值不变直到Re 发生变化？

我的数据框：

    Dates()  Close  Re  Size
0  12/29/10  86.76   0  2073
1  12/30/10  86.54   1  2209
2  12/31/10  86.30   1  2339
3  08/09/11  82.69   0   984
4  08/10/11  78.23   0   880
5  09/02/11  79.34  -1  1015
6  09/06/11  78.08  -1  1002

Answer 1

你可以使用loc:

print df
    Dates()  Close  Re  Size
0  12/29/10  86.76   0  2073
1  12/30/10  86.54   1  2209
2  12/31/10  86.30   1  2339
3  08/09/11  82.69   0   984
4  08/10/11  78.23   0   880
5  09/02/11  79.34  -1  1015
6  09/06/11  78.08  -1  1002

df.loc[df['Re']==0, 'Position'] = 0
df.loc[df['Re']==1, 'Position'] = df['Size']
df.loc[df['Re']==-1, 'Position'] = -df['Size']

print df
    Dates()  Close  Re  Size  Position
0  12/29/10  86.76   0  2073         0
1  12/30/10  86.54   1  2209      2209
2  12/31/10  86.30   1  2339      2339
3  08/09/11  82.69   0   984         0
4  08/10/11  78.23   0   880         0
5  09/02/11  79.34  -1  1015     -1015
6  09/06/11  78.08  -1  1002     -1002

编辑：

您可以将fillna 与method='ffill' 一起使用：

print df
     Dates()  Close  Re  Size  Position
0   12/30/10  86.54   1   100       100
1   12/30/10  86.54   1   100       100
2   12/29/10  86.76   0  2073         0
3   12/30/10  86.54   1  2209      2209
4   12/30/10  86.54   1 -2209     -2209
5   12/29/10  86.76   0  2073         0
6   12/30/10  86.54   1  2209      2209
7   12/29/10  86.76   0  2073         0
8   12/31/10  86.30   1  2339      2339
9   08/09/11  82.69   0   984         0
10  08/10/11  78.23   0   880         0
11  09/02/11  79.34  -1  1015     -1015
12  09/06/11  78.08  -1  1002     -1002
13  09/06/11  78.08   1    20        20


df.loc[((df['Position'].shift()).fillna(1)) == 0,'pos'] = df['Position']
df.loc[ df['Position'] == 0, 'pos'] = 0
df.at[0,'pos'] = df.at[0,'Position']
df['pos'] = df['pos'].fillna(method='ffill')

print df
     Dates()  Close  Re  Size  Position   pos
0   12/30/10  86.54   1   100       100   100
1   12/30/10  86.54   1   100       100   100
2   12/29/10  86.76   0  2073         0     0
3   12/30/10  86.54   1  2209      2209  2209
4   12/30/10  86.54   1 -2209     -2209  2209
5   12/29/10  86.76   0  2073         0     0
6   12/30/10  86.54   1  2209      2209  2209
7   12/29/10  86.76   0  2073         0     0
8   12/31/10  86.30   1  2339      2339  2339
9   08/09/11  82.69   0   984         0     0
10  08/10/11  78.23   0   880         0     0
11  09/02/11  79.34  -1  1015     -1015 -1015
12  09/06/11  78.08  -1  1002     -1002 -1015
13  09/06/11  78.08   1    20        20 -1015