Mathematical calculation using for loop in a dataframe

I have a dataframe which looks like this

df= 

    Time                  x          y
0   2018-09-13 01:17:00  5.0        0.0
1   2018-09-13 02:17:00  9.0        0.0
2   2018-09-13 03:17:00  2.0        1.0
3   2018-09-13 04:17:00  1.0        0.0

…….
I want to iterate through this whole dataframe and calculate a new variable z.
The value of z would be z= z[prev]+ x-y

for example, the final output would be

Time                         z
0   2018-09-13 01:17:00      5    #[0+5-0]
1   2018-09-13 02:17:00      14   #[5+9-0]
2   2018-09-13 03:17:00      15   #[14+2-1]
3   2018-09-13 04:17:00      16   #[15+1-0]

…..

I am finding it difficult to iterate over the time series data.

I have tried the following but it is not working.

for i,row in df.iterrows():
    z=0
    row['z']=row['z']+row['x']-row['y']

print[z]

>Solution :

In your case do cumsum

df['new'] = df.x.sub(df.y).cumsum()
Out[410]: 
0  2018-09-13     5.0
1  2018-09-13    14.0
2  2018-09-13    15.0
3  2018-09-13    16.0
dtype: float64