Field Sorting

Summary

• By default, ft.fields provides a list sorted by column number in ascending order.
• If you passed a sequence of desired fields (FuzzyTable(fields=...)), the ft.fields fields will be sorted in the same order.
• MultiField subfields are sorted by column number

Details

Let’s read the following csv file.

first_name	last_name	birthday
John	Doe	3-Mar-85
Typhoid	Mary	2-Aug-83
Jane	Smith	23-Oct-46

Let’s extract this table using all the default settings. Notice how ft.fields returns the three fields in the same order they appear in the spreadsheet.

>>> import fuzzytable
>>> ft = fuzzytable.FuzzyTable('birthdays.csv')
>>> [field.name for field in ft.fields]
['first_name', 'last_name', 'birthday']

But now let’s extract the table using the field parameter. Notice now how ft.fields returns the fields in the order as passed to FuzzyTable.

>>> ft = fuzzytable.FuzzyTable(
>>>     'birthdays.csv',
>>>     fields=['last name', 'first name'],
>>> )
>>> [field.name for field in ft.fields]
['last_name', 'first_name']
>>>

But what if you don’t want them in that order? What if you’d prefer to have them sorted in the order they appear in the spreadsheet? Easy. Because ft.fields is a list, you can sort it in place. See https://docs.python.org/3/howto/sorting.html for more details.

>>> from operator import attrgetter
>>> ft.fields.sort(key=attrgetter('col_num'))
>>> [field.name for field in ft.fields]
['first_name', 'last_name']

The subfields collected by a multifield=True FieldPattern are stored in spreadsheet order. The col_num of a multifield FieldPattern is equal to the smallest col_num of its subfields.

>>> names = fuzzytable.FieldPattern(
>>>     name='names',
>>>     alias=['last name', 'first name'],
>>>     multifield=True,
>>>     approximate_match=True,
>>> )
>>>
>>> ft = fuzzytable.FuzzyTable(
>>>     'birthdays.csv',
>>>     fields=[names, 'birthday'],
>>> )
>>>
>>> [field.name for field in ft.fields]
['names', 'birthday']
>>> namesfield = ft.get_field('names')
>>> namesfield.headers
['first_name', 'last_name']
>>> # Note that these are in their original table order, not the alias order.