header-detail-footer 2.2

Overview

The header_detail_footer module provides a way to parse input iterables (usually text files) that contain header rows, an unknown number of data rows (called the details), and footer rows. The number of header and footer rows must be specified when parsing begins. There can be zero or more header rows, and independently zero or more footer rows. The detail rows are all of the rows between the header and footer. If a file consists of just headers and footers, there will be zero detail rows.

The module API consists of a single function, parse(), and several exceptions.

Note that the contents of each input “row” are never inspected: they are just iterated over and returned by the parser. They are often strings, but they could be any object.

Typical usage

This code shows a simple usage of the parse() function:

>>> from header_detail_footer import parse
>>> header, details, footer = parse(['header', 'row 1', 'row 2', 'footer'])
>>> header
'header'
>>> list(details)
['row 1', 'row 2']
>>> footer()
'footer'

The parse() function

The parse() function takes 1 required parameter, the input iterable. There are two optional parameters, header_rows and footer_rows. Both default to 1. They represent the number of header and footer rows present in the input, respectively.

parse() returns a 3-tuple: (header, details, footer). header is the header row(s), if any; details is an iterator returning each detail row; and footer is a callable returning the footer row(s), if any.

For header and footer(), they return a single row from the input if header_rows or footer_rows is 1, respectively. Otherwise, including the case of 0 rows, they contain a list:

>>> header, details, footer = parse(['row 1', 'row 2', 'footer 1', 'footer 2'],
...                                 header_rows=0, footer_rows=2)
>>> header
[]
>>> list(details)
['row 1', 'row 2']
>>> footer()
['footer 1', 'footer 2']

The returned footer callable need never be called. If footer is ever called, details must have been exhausted, otherwise a RuntimeError is raised:

>>> header, detail, footer = parse('abc')
>>> footer()
Traceback (most recent call last):
    ...
RuntimeError: called footer() before details were exhausted

Exceptions

>>> header, details, footer = parse(['row 1'], header_rows=3)
Traceback (most recent call last):
    ...
HeaderError: too few rows for header

>>> header, details, footer = parse(['row 1', 'row 2', 'row 3'],
...                                 header_rows=2, footer_rows=2)
Traceback (most recent call last):
    ...
FooterError: too few rows for footer

Usage with CSV files

If you have a CSV file with a header row containing column names, and also with footer row, it’s easiest if you tell header_detail_footer that there is no header row. That way, the details iterator can be passed to csv.DictReader:

>>> import csv

# here, the footer row contains a row count
>>> _, details, footer = parse(['FRUIT,COLOR',
...                             'apple,red',
...                             'orange,orange',
...                             'rowcount:2'],
...                            header_rows=0)  # specify 0 header rows

# pass the details to csv.DictReader. this includes what is now
# the csv header row
>>> reader = csv.DictReader(details)

# print out each row
>>> for count, row in enumerate(reader, 1):
...      [(key, value) for key, value in sorted(row.items())]
...
[('COLOR', 'red'), ('FRUIT', 'apple')]
[('COLOR', 'orange'), ('FRUIT', 'orange')]

# verify the footer count
>>> _, _, footer_count = footer().partition(':')
>>> int(footer_count) == count
True

Change log