sqlobject.util.csvimport module

Import from a CSV file or directory of files.

CSV files should have a header line that lists columns. Headers can also be appended with :type to indicate the type of the field. escaped is the default, though it can be overridden by the importer. Supported types:

:python:

A python expression, run through eval(). This can be a security risk, pass in allow_python=False if you don’t want to allow it.

:int:

Integer

:float:

Float

:str:

String

:escaped:

A string with backslash escapes (note that you don’t put quotation marks around the value)

:base64:

A base64-encoded string

:date:

ISO date, like YYYY-MM-DD; this can also be NOW+days or NOW-days

:datetime:

ISO date/time like YYYY-MM-DDTHH:MM:SS (either T or a space can be used to separate the time, and seconds are optional). This can also be NOW+seconds or NOW-seconds

:bool:

Converts true/false/yes/no/on/off/1/0 to boolean value

:ref:

This will be resolved to the ID of the object named in this column (None if the column is empty). @@: Since there’s no ordering, there’s no way to promise the object already exists.

You can also get back references to the objects if you have a special [name] column.

Any column named [comment] or with no name will be ignored.

In any column you can put [default] to exclude the value and use whatever default the class wants. [null] will use NULL.

Lines that begin with [comment] are ignored.

sqlobject.util.csvimport.load_csv_from_directory(directory, allow_python=True, default_type='escaped', allow_multiple_classes=True)

Load the data from all the files in a directory. Filenames indicate the class, with general.csv for data not associated with a class. Return data just like load_csv does.

This might cause problems on case-insensitive filesystems.

sqlobject.util.csvimport.load_csv(csvreader, allow_python=True, default_type='escaped', default_class=None, allow_multiple_classes=True)

Loads the CSV file, returning a list of dictionaries with types coerced.

sqlobject.util.csvimport.create_data(data, class_getter, keyorder=None)

Create the data, which is the return value from load_csv(). Classes will be resolved with the callable class_getter; or if class_getter is a module then the class names will be attributes of that.

Returns a dictionary of {object_name: object(s)}, using the names from the [name] columns (if there are any). If a name is used multiple times, you get a list of objects, not a single object.

If keyorder is given, then the keys will be retrieved in that order. It can be a list/tuple of names, or a sorting function. If not given and class_getter is a module and has a soClasses function, then that will be used for the order.