a = {'A':[1,2,5,7,0,9],
     'B':[6,5,0,0,7,2],
     'C':[0,8,np.nan,10,0,6],
     'D':[np.nan, 9,5,2,6,7],
     'E':[1,4,6,3,3,6]}
aidx = ['id_1','id_2','id_3',
        'id_4','id_5','id_6']
df = pd.DataFrame(a, index=aidx)

def get_col(df,num):
    df_num = df==num
    df_num=df_num[df_num.any(axis=1)].idxmax(axis=1)
    return(df_num)
df_new = pd.DataFrame(get_col(df,0))

df_need = pd.DataFrame([2,999,1,1,0,999], index=aidx)