Length of first sequence of zeros of given size after certain column in pandas dataframe

Question

Suppose I have a dataframe like this:

        ID      0   1   2   3   4   5   6   7   8   ... 81  82  83  84  85  86  87  88  89  90  total  day_90
----------         


        

Answer 1

Here is my solution, see the comments in the code:

import numpy as np, pandas as pd
import io

# Test data:
text="""  ID  0   1   2   3  4  5   6   7  8  day_90
        0  A  2  21   0  18  3  0   0   0  2       4
        1  B  0  20  12   2  0  8  14  23  0       5
        2  C  0  38  19   3  1  3   3   7  1       1
        3  D  3   0   0   1  0  0   0   0  0       0"""

df= pd.read_csv( io.StringIO(text),sep=r"\s+",engine="python")
#------------------------

# Convert some column names into integer:
cols= list(range(9))
df.columns= ["ID"]+ cols +["day_90"]

#----------
istart,istop= df.columns.get_loc(0), df.columns.get_loc(8)+1
# The required length of the 1st zero sequence:
lseq= 2

# The function for aggregating: this is the main calculation, 'r' is a row of 'df':
def zz(r):

     s= r.iloc[r.day_90+istart:istop] # get the day columns starting with as fixed in 'day_90'
     #--- Manipulate 's' to make possible using 'groupby' for getting different sequences:
     crit=s.eq(0)
     s= pd.Series(np.where(crit, np.nan, np.arange(len(s))),index=s.index)
     if np.isnan(s.iloc[0]):
       s.iloc[0]= 1
     s= s.ffill()
     s[~crit]= np.nan
     #---
     # get the sequences and their sizes:
     ssiz= s.groupby(s).size()
     return ssiz.iloc[0] if len(ssiz) and ssiz.iloc[0]>lseq else np.nan
#---

df["zseq"]= df.agg(zz,axis=1)

ID  0   1   2   3  4  5   6   7  8  day_90  zseq
0  A  2  21   0  18  3  0   0   0  2       4   3.0
1  B  0  20  12   2  0  8  14  23  0       5   NaN
2  C  0  38  19   3  1  3   3   7  1       1   NaN
3  D  3   0   0   1  0  0   0   0  0       0   NaN