Python数据框总和行

Question

If I have a dataframe with n rows, is there a way to set the ith row to be sum of row[i] and row[i-1] and do this so that the assignment to earlier rows is reflected in the latter rows? I would really like to avoid loops if possible.

Example DF:

             SPY   AAPL   GOOG
2011-01-10  0.44  -0.81   1.80
2011-01-11  0.00   0.00   0.00
2011-01-12 -1.11  -2.77  -0.86
2011-01-13 -0.91  -4.02  -0.68

Sample pseudo code of summing two rows:

DF[2011-01-11] = DF[2011-01-10] + DF[2011-01-11]
DF[2011-01-12] = DF[2011-01-11] + DF[2011-01-12]

and so on.

Answer 1

based on your question, you are looking for a cumulative sum of each columns. you could use the cumsum() method

DF.cumsum()

               SPY    AAPL   GOOG
2011-01-10  0.4400 -0.8100 1.8000
2011-01-11  0.4400 -0.8100 1.8000
2011-01-12 -0.6700 -3.5800 0.9400
2011-01-13 -1.5800 -7.6000 0.2600

Answer 2

使用DF.shift()

DF + DF.shift()