python pandas 嵌套循环：将函数应用于例如第 2 列的每个元素，涉及复合前一列中的相同元素

Question

I am new to pandas/data frame programming, and have not yet grasped implementing the equivalent of simple loops.

What I am essentially trying to implement is something like this:

counter = 0
for port in portfolios:
    for bin in bins:
        value[port[counter +1, bin]] 
        = port[counter + 1, bin] * port[counter, bin]
counter += 1

In fact, using an example table below, for each bin in each portfolio port, I am trying to multiply rebal_wgt [port=i,bin=j] by rebal_wgt[port=i-1,bin=j] . I want to do this for all portfolios port (there are 11 of them), and all the bins within each portfolio port, such that each portfolio's corresponding bin gets multiplied by the previous portfolio's corresponding bin (eg port5 bin 1 rebal_weight * port 4 bin 1's rebal_weight etc.)

---

In fact, I think that what I am trying to achieve is a cumulative product.
From above,

value[port[counter +1, bin]] = port[counter + 1, bin]*port[counter, bin]

So it stands to reason that any value[port,bin] can be expressed as a product with the value[port-1,bin]. Thus, starting with value[0,bin], one can do cumulative product to derive the values of successive portfolios for each bin. Which is what I would like to do! unfortunately I don't think the first answer given below does this as it doesn't loop over portfolios for each bin, it seems to loop over bins within each portfolio. Any help much appreciated! a sample data table is below, ignore cum_ret_sum and stocks.

     port  bin   cum_ret_sum  stocks     rebal_wgt  
  0       0    0  4.067563e+03     216  1.883131e+01  
  1       0    1  1.300282e+04     213  6.104612e+01  
  2       0    2  1.426061e+04     214  6.663837e+01  
  3       0    3  4.904957e+02     205  2.392662e+00  
  4       0    4  1.100993e+04     209  5.267908e+01  
  5       0    5  4.630904e+03     208  2.226396e+01  
  6       0    6  1.019425e+04     215  4.741514e+01  
  7       0    7  2.249585e+04     213  1.056143e+02  
  8       0    8  8.831653e+03     214  4.126941e+01  
  9       0    9  3.098015e+05     212  1.461328e+03  
  10      1    0  1.881155e+00     267  7.045525e-03  
  11      1    1  7.486650e+00     280  2.673804e-02  
  12      1    2  4.492010e+00     268  1.676123e-02  
  13      1    3  1.191500e+01     273  4.364468e-02  
  14      1    4  4.388776e+00     266  1.649916e-02  
  15      1    5  1.384601e+01     270  5.128153e-02  

Example output for bin = 0 (want this for all bin = 0 to bins)

port    bin rebal_wgt   rebal_wgt(port[counter,bin] * port[counter - 1, bin])
0   0   18.83131173675383   18.83131173675383
0   1   61.04612379316069   
0   2   66.63837376523843   
0   3   2.3926618103462602  
0   4   52.67907609709447   
0   5   22.26396127204363   
0   6   47.415137683181634  
0   7   105.61432751079496  
0   8   41.26940808193825   
0   9   1461.3276714958988  
1   0   0.0070455247063908755   0.13267647209504757
1   1   0.026738035817712412    
1   2   0.01676122975460385 
1   3   0.04364467915381678 
1   4   0.016499159440430355    
1   5   0.051281526270788164    
1   6   0.04977016623588389 
1   7   0.0645014820724396  
1   8   0.1438106018214078  
1   9   0.0340451076286303  
2   0   5.1196753262692285  0.6792604605614628
2   1   0.014870173557215314    
2   2   3.263374203937453   
2   3   73.32640040253595   
2   4   3.915173886409575   
2   5   67.46028895344207   
2   6   7.654613865824991   
2   7   12.837204120226547  
2   8   2983.065107673766   
2   9   0.4204701203425892  
3   0   71.16936245719319   48.34253392053873

Actual desired output:

port    bin rebal_wgt   rebal_wgt(port[counter,bin] * port[counter - 1, bin])
0   0   18.83131173675383   18.83131173675383
0   1   61.04612379316069   61.04612379316069
0   2   66.63837376523843   66.63837376523843
0   3   2.3926618103462602  2.3926618103462602
0   4   52.67907609709447   52.67907609709447
0   5   22.26396127204363   22.26396127204363
0   6   47.415137683181634  47.415137683181634
0   7   105.61432751079496  105.61432751079496
0   8   41.26940808193825   41.26940808193825
0   9   1461.3276714958988  1461.3276714958988
1   0   0.0070455247063908755   0.13267647209504757
1   1   0.026738035817712412    1.6322534445140364
1   2   0.01676122975460385 1.116941093152327
1   3   0.04364467915381678 0.10442695703615294
1   4   0.016499159440430355    0.8691604757005253
1   5   0.051281526270788164    1.1417299148641158
1   6   0.04977016623588389 2.3598592845892727
1   7   0.0645014820724396  6.812280652530306
1   8   0.1438106018214078  5.934978413076811
1   9   0.0340451076286303  49.75105785677358
2   0   5.1196753262692285  0.6792604605614628
2   1   0.014870173557215314    0.024271892009286238
2   2   3.263374203937453   3.6449967507110035
2   3   73.32640040253595   7.65725286445137
2   4   3.915173886409575   3.4029143975620206
2   5   67.46028895344207   77.02142996352207
2   6   7.654613865824991   18.06381160121289
2   7   12.837204120226547  87.45063726080163
2   8   2983.065107673766   17704.427018846454
2   9   0.4204701203425892  20.918833284208702

Answer 1

cumulative operation

---

In [5]:
df.groupby(df.bin)['rebal_wgt'].cumprod()
Out[5]:
0       18.831310
1       61.046120
2       66.638370
3        2.392662
4       52.679080
5       22.263960
6       47.415140
7      105.614300
8       41.269410
9     1461.328000
10       0.132676
11       1.632254
12       1.116941
13       0.104427
14       0.869161
15       1.141730
dtype: float64