dataframe: forward fill next n rows on condition-CodePudding

I have a dataframe that looks like this

df = pd.DataFrame({'Col1':[0,1,0,1,0,1,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0]})

I want to have a rolling function where if the value of 1 appears, the next 5 rows will flip to 0.

I think it'll be have to an iterative process, but not sure how best to implement it.

This is the desired df using the above rule.

df_desired = pd.DataFrame({'Col1':[0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0]})

Thanks

CodePudding user response：

So to do this, you'll want to find all the indices of df that are equal to 1 and then find all the indices that are greater than 5 apart from each other consecutively.

df = pd.DataFrame({'Col1':[0,1,0,1,0,1,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0]})
ind = df[df['Col1'] == 1].index #get the indices where df == 1

#find all the indices that are consecutively greater than 5 apart
val = [ind[0]]
for k in ind[1:]:
    if k - val[-1] > 5:
        val.append(k)

df['Col1'] = 0 #change all the values to zero
df['Col1'][val] = 1 #change all indices where there was spaced ones to 1

CodePudding user response：

Here is a simple solution:

i = 0
while i < (len(list) - 5):

    if list[i] == 1:
        list[i 1:i 6] = [0]*5
        i  = 6

    i  = 1
print(list)

You just replace the slice of the next 5 values with 0's. ([0]*5 creates such an array). Then you skip by 6 if you replaced, otherwise just by the usual 1.

Of course you need to convert to a list first using pd.Series.tolist(). Then you can add it back

CodePudding user response：

Another solution to what has already been provided. This just iterates through the data frame. Each time it encounters a 1, it will start flipping the next 5 elements to 0.

import pandas as pd
no_of_flips = 5
df = pd.DataFrame({'Col1':[0,1,0,1,0,1,1,0,1,0,0,0,1,0,0,1,0,1,0,0,0]})
df_desired = pd.DataFrame({'Col1':[0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0]})

for inx in df.index:
    if i > 0:
        df.at[inx, 'Col1'] = 0
        i -= 1
    if df.loc[inx,'Col1'] == 1:  
        i = no_of_flips

if (df.compare(df_desired)).empty:
    print('Both dataFrames are the same!')
else:
    print('Dataframes are different')