How to keep consecutive rows based on values of two different columns in datatframe-CodePudding

I have a dataframe :

        Blocks  Response        RT     Name  
1       7       True            4630   ya__79891       
2       7       True            4610   ya__79891             
3       6       True            4390   ya__79891            
4       6       True            5190   ya__79891            
5       6       True            4260   ya__79891             
6       5       True            3560   ya__79891           
7       5       True            3610   ya__79891

I want to keep the latest or top-most rows which have same Blocks values and Response as True consecutively for 3 times i.e. in the above example keep rows 3,4,5 only and remove all the others.

the desired output :

        Blocks  Response        RT     Name             
3       6       True            4390   ya__79891            
4       6       True            5190   ya__79891            
5       6       True            4260   ya__79891

Is there any shorter method to do this? here is an incomplete code of what I have done :

df_1['Points'] = df_1['Response'].astype(int) #converted all the True and False values as 0 and 1
df_1 = df_1[df_1.Points != 0] #remove all the False/0 values

tmp = list(df_1['Blocks']) #created a list of the dataframe colum blocks
tmp = dict(Counter(tmp)) #created a dict`]
wm = [] #empty list for storing Blocks values`
for key,value in tmp.items():
        if(value==3):
            mem = key
            break
    wm.append(mem)
df2 = pd.DataFrame() # new data frame for saving values
df2['Name'] = name
df2['span'] = wm

Am still not able to get everything that I need as given above using this code.

Can someone help ?

CodePudding user response：

Try this:

df[df.groupby(['Blocks', 'Response'])['RT'].transform('size') > 2]

Output:

   Blocks  Response    RT       Name
3       6      True  4390  ya__79891
4       6      True  5190  ya__79891
5       6      True  4260  ya__79891