csvfile

csvfile

A simple way of working with csv files.

You can use:

data = csvfile.load("my-data.csv")
data[0].field = "new value"
data.sync()

Instead of:

data = []
with open("my-data.csv") as f:
    reader = csv.DictReader(f)
    fieldnames = reader.fieldnames
    for row in reader:
        data.append(row)

data[0]["field"] = "new value"

with open("my-data.csv, "w") as f:
    writer = csv.DictWriter(f, fieldnames=fieldnames)
    writer.writeheader()
    writer.writerows(data)

Also you can declaratively define types in a table's header:

>>> print(open("/tmp/csvfile_test.csv").read())
language,created:i
python,1991
js,1995
rust,2010

>>> pprint(csvfile.load("/tmp/csvfile_test.csv"))
[{'created': 1991, 'language': 'python'},
 {'created': 1995, 'language': 'js'},
 {'created': 2010, 'language': 'rust'}]

Notice, that created is automatically converted into integer as we typed it in the header as created:i.

Built-in types

Type	Alias	Comment
str	s	default, if no type specified
bool	b	reader can understand next pairs as True / False case-insensitevely: true / false, t / f, 1 / 0, y / n, yes / no, on / off, but the writer will always stringify them into true / false
int	i
float	f
decimal	n
date	d	ISO 8601, YYYY-MM-DD
datetime	t	ISO 8601, YYYY-MM-DDTHH:MM:SS.ffffff
json	j

Empty cells will become None for any type except of str. As in case of str there's no way to distinguish it from an empty string.

API

The api consists in a single function load with the signature:

table = load(
    filename: Union[Path, str],   # the file should exist and contain headers
    *csvargs: Any,                # `*args` for the standard `csv.reader`
    limit: Optional[int] = None,  # maximum number of rows to read
    **csvkwargs: Any,             # `**fmtparams` for the standard `csv.reader`
)

The returned object has a list interface with an additional sync method to save changes back into the file.

Running tests

pip install -r requirements-dev.txt
pytest