Parse textfile without fixed structure using python dictionary and Pandas - python-2.7

Question

I have .txt file without specific separators and to parse it, I need to count character by character to know where starts and ends a column. To do so, I constructed a Python dictionary where the keys are the column names and the values are the number of characters that takes each column:
headers = {first_col: 3, second_col: 5, third_col: 2, ... nth_col: n_chars}
Having that in mind, I know that the three first columns of the following line in the .txt file
ABC123-3YN0000000001203ABC123*TESTINGLINE
first_col: ABC
second_col: 123-3
third_col: YN
I want to know if there is any pandas function that helps me to parse this .txt taking into account this particular condition and (if possible) using my headers dictionary.

Answer 1

Using a dictionary is dangerous because the order is not guaranteed. Meaning, if you picked third_col first, you've thrown of your entire scheme. You can fix this by using lists. From there, you can use pd.read_fwf to read a fixed with formatted text file.
Solution
names = ['first_col', 'second_col', 'third_col']
widths = [3, 5, 2]
pd.read_fwf(
'myfile.txt',
widths=widths,
names=names
)
first_col second_col third_col
0 ABC 123-3 YN
You can also use OrderedDict from the collections library and make sure you keep the order you want by passing an iterator that produces tuples in the correct order
from collections import OrderedDict
names = ['first_col', 'second_col', 'third_col']
widths = [3, 5, 2]
header = OrderedDict(zip(names, widths))
pd.read_fwf(
'myfile.txt',
widths=header.values(),
names=header.keys()
)
first_col second_col third_col
0 ABC 123-3 YN
Demonstration
from collections import OrderedDict
txt = """ABC123-3YN0000000001203ABC123*TESTINGLINE"""
names = ['first_col', 'second_col', 'third_col']
widths = [3, 5, 2]
header = OrderedDict(zip(names, widths))
pd.read_fwf(
'myfile.txt',
widths=header.values(),
names=header.keys()
)
first_col second_col third_col
0 ABC 123-3 YN