Drop rows where a certain column has more than n zeros in a row

I am processing data from a datalogger which collects data at millisecond intervals, regardless of whether anything interesting is happening. As a result, there are long stretches of data where certain columns are zero. I would like to drop these rows, but leave some number of them as padding, which will help in further processing.

Example input:

a b c
1 2 3
0 3 2
0 1 0
0 3 7
0 9 6
4 0 1
0 2 5
0 6 3
0 1 6

Example output - filter on column a, replace column with more than 2 zero rows with exactly 2 zero rows (selection of which rows to drop can be arbitrary):

a b c
1 2 3
0 3 2
0 1 0
4 0 1
0 2 5
0 6 3

My current code drops all rows where a column is zero:

df= df[df['ColName'] != 0]

You can use a rolling (n+1) sum to exclude the rows you wouldn't like to keep. For example for two consecutive 0s you can use:

exclude = df['a'].rolling(3).sum() != 0

And then:

output = df.loc[exclude,:]

Returning:

   a  b  c
0  1  2  3
1  0  3  2
2  0  1  0
5  4  0  1
6  0  2  5
7  0  6  3

Or, in a one-liner:

output = df[df['a'].rolling(3).sum() != 0]

You can change the number in the rolling function to fit the amount of periods you'd like to check.