Pandas Series: How to get the nth non-zero element before each element

Question

I have a pandas series like this:

[0,3,4,5,0,0,2,4,5]

For each element, I want to be able to get the nth non-zero element before it.

The output Series for the above input Series would be

n=1
[NaN,NaN,3,4,5,5,5,2,4]

n=2
[NaN,NaN,NaN,3,4,4,4,5,2]

etc.

Is there a built in way in pandas to do this easily?

Answer 1

You can use rolling and apply to extract the nth non zero elements from the rolling window.

First we use rolling(len(s)) which gives a rolling 9 rows window until the current row for each row. Then we remove 0s from the rolling window. x[:-1] is to exclude the current row because we need to find the non-zero element before it. Finally, we extract the nth element from the non-zero list by doing [-n].

n=1
s.rolling(len(s),min_periods=0)\
 .apply(lambda x: x[:-1][x[:-1]!=0][-n] if len(x[:-1][x[:-1]!=0])>=n else np.nan)
Out[97]: 
0    NaN
1    NaN
2    3.0
3    4.0
4    5.0
5    5.0
6    5.0
7    2.0
8    4.0
dtype: float64

n=2    
s.rolling(len(s),min_periods=0)\
 .apply(lambda x: x[:-1][x[:-1]!=0][-n] if len(x[:-1][x[:-1]!=0])>=n else np.nan)
Out[99]: 
0    NaN
1    NaN
2    NaN
3    3.0
4    4.0
5    4.0
6    4.0
7    5.0
8    2.0
dtype: float64

Answer 2

def dew(zoop, n):
  x = zoop.pop()
  zoop.reverse()
  for i in range(len(zoop)):
    if not zoop[i] == 0:
      n -= 1
      if n == 0:
        return zoop[i]
  return float('NaN')

lzt = [0,3,4,5,0,0,2,4,5]
ans = []
for i in range(len(lzt)):
  ans.append(dew(lzt[:i+1], 2)) #2 is n here