I know this subject is already posted but I still don't understand the windows function in pyspark. I just want to do this on a pyspsark dataframe : data.rolling(5).agg('sum') -> this is in Pandas.

I want it in pyspark. No need to groupby or orderby, just slide a window on a column and calcul the sum (or my own function).

Example :

df = pd.DataFrame({'A': [1,1,2,2,1,2],
                    'B': [2,2,3,4,2,1]})

print(df)
   A  B
0  1  2
1  1  2
2  2  3
3  2  4
4  1  2
5  2  1

Result :

print(df.rolling(3).agg('sum'))
     A    B
0  NaN  NaN
1  NaN  NaN
2  4.0  7.0
3  5.0  9.0
4  5.0  9.0
5  5.0  7.0

Thanks

You can achieve this by creating a single window and limiting rows to aggregate

from pyspark.sql import Window
from pyspark.sql.functions import *


df1.show()
+---+---+
| v1| v2|
+---+---+
|  1|  2|
|  1|  4|
|  2|  2|
|  2|  4|
|  2|  4|
|  2|  4|
|  2|  4|
|  2|  4|
+---+---+


w = Window().partitionBy(lit(1)).rowsBetween(-2,0)
df1.select(sum('v1').over(w).alias('v1'),sum('v2').over(w).alias('v2')).show()

+---+---+
| v1| v2|
+---+---+
|  1|  2|
|  2|  6|
|  4|  8|
|  5| 10|
|  6| 10|
|  6| 12|
|  6| 12|
|  6| 12|
+---+---+

You can explicitly set first two rows null if you want