Load (double) structured txt file with pandas

Question

I try to load a public txt file into a dataframe in Pandas to execute a Name Entity recognition in the german language later. the original txt file has the structure # words [date ] followed by a number (Position in the sentence), a word and the Name entitiy recognition and the words are seperated with Tabs. So the structure is:

text [21-03-1991] 1 Aufgrund O O 2 des O O # text [22-04-1993] 1 Aber O P

has anyone an idea how can i get it into this format:

Aufgrund 0 0
des 0 0

Aber O P

best case every # in a new column?

i would like to use

pd.read_csv(...)

Answer 1

Text file example

text  [21-03-1991]    1   Aufgrund    O   O   2   des O   O   #   text    [22-04-1991]    1   Aber    O   P
text  [21-04-1992]    2   Aufgrund    O   O   3   des O   O   #   text    [22-04-1992]    1   Aber    O   P
text  [21-06-1993]    3   Aufgrund    O   O   5   des O   O   #   text    [22-04-1993]    1   Aber    O   P

import pandas as pd

# Reading tab separated text file
df = pd.read_csv("source.txt",sep='\t')

df1 = df.iloc[:,[3,4,5]]
df1.columns = ['V1','V2','V3']
df2 = df.iloc[:,[7,8,9]]
df2.columns = ['V1','V2','V3']
df3 = df.iloc[:,[14,15,16]]
df3.columns = ['V1','V2','V3']

d_one = df1.append(df2, ignore_index=True)
final_df = d_one.append(df3, ignore_index=True)
print(final_df)