Pandas next row minus previous row

1.diff() function df['1].diff()
The first result is NAN.

2.shift() function ：df['1'].shift(-1) - df['1'] Represents the next line minus the previous line . The last line of the result is NAN.

• pandas DataFrame.shift() The function can move the data by a specified number of bits .
• period Parameter specifies the stride of the movement , Can be positive or negative .axis Specifies the axis to move ,1 For the line ,0 Column .

Series.shift(periods=1, freq=None, axis=0, fill_value=None)

shift() function ： Default down （axis=0） To the right （axis=1）.
If you want to move up and left , You need to set the parameters periods=-1, That is, move up ：data.shift(periods=-1, axis=0)
Move to the left ：data.shift(periods=-1, axis=1)

3.pandas Single column 、 Multiple columns for calculation
Perform formula calculation for multiple columns ：
https://blog.csdn.net/zwhooo/article/details/79696558

4.python Middle up and down 、 Operation between columns

Change the line to the previous line ：

df['B'] = math.sqrt(df['A'] * df['A'].shift(periods=-1, axis=0))

Report errors ：

terms of settlement ：https://stackoverflow.com/questions/42988348/typeerror-cannot-convert-the-series-to-class-float

Change the code to ：

df["B"] = (df["A"] * df["A"].shift(periods=-1, axis=0)).apply(lambda x: math.sqrt(x))

5. Calculate the cumulative sum
cumsum() function ： Seek before n Cumulative value of elements

df["B"] = df['A'].cumsum(axis=0) # Calculated by line