how to conditionally remove rows from a pandas expanding window

I have a series in which I want to take the cumulative median of all non-zero values, resulting in a series the same length as the original.

my_series.expanding().median() gives me a series the same length as my_series which is close to what I want, but before I take the median of each window I want to drop rows that equal zero from the window, or slice out non-zero values, or something else… whatever performs best.

a = [0, 1, 2, 0, 100, 1000]
my_series = pd.Series(a)

my_series.expanding().median()

# returns:

0    0.0
1    0.5
2    1.0
3    0.5
4    1.0
5    1.5
dtype: float64

# desired output:
# the median is only computed on values in each window that are greater than zero

0    nan
1     1.0
2     1.5
3     1.5
4     2.0
5    51.0
dtype: float64

>Solution :

You can replace 0 values with nan while calculating, so they won’t be used in the median calculations.

 my_series.replace(0, np.nan).expanding().median()

Output:

0     NaN
1     1.0
2     1.5
3     1.5
4     2.0
5    51.0
dtype: float64