Django-Adaptors TypeError：“定界符”必须为1个字符的字符串

Question

I am getting an error when trying to import some data into my model. The error I'm getting is TypeError: complaining about the delimiter I'm using.

Below is my model for the CSV import, I am using the default delimiter suggested by the documentation.

class SkuCsvModel(CsvModel):
sku_num = models.CharField()
sku_category = models.ForeignKey(SkuCategory)
short_desc = models.CharField()

class Meta:
    delimiter = ";"
    dbModel = Sku

The CSV file I'm trying to use is below:

 1365400;9;3/8 BALL VALVE
 1401901;9;BRASS ELBOW
 1406300;9;HOSE BARB, NPT

The code I'm testing in the manage.py shell is:

>>> from core.models import SkuCsvModel
>>> my_csv_list = SkuCsvModel.import_data(data = open("labconco.csv"))

And finally the error I'm getting is:

Traceback (most recent call last):
  File "<console>", line 1, in <module>
  File "E:\bin\Python27\lib\site-packages\adaptor\model.py", line 197, in import_data
    return importer.import_data(data)
  File "E:\bin\Python27\lib\site-packages\adaptor\model.py", line 466, in import_data
    for line in csv.reader(data, delimiter=self.delimiter):
TypeError: "delimiter" must be an 1-character string

So I've been fiddling around with the django-adaptor tools, and this error is coming from the import_data() method of the CsvImporter, when I try and put a delimiter directly into the csv.reader(data, delimiter=';') this works fine and I'm able to see the file correctly. But no matter how I try and enter this import_data method sending in a ';' will generate an error.

Answer 1

Look at the snippet below. If I provide an integer as a delimiter, it fails with the same exception as in your example. If I provide a semicolon as a delimiter to csv.reader it works. This is basically what is done in model.CsvImporter.import_data() as you already found out.

>>> import csv
>>> import StringIO
>>> io = StringIO.StringIO('name;surname\nsascha;gottfried')
>>> for line in csv.reader(io, delimiter=10):
...     print line
... 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: "delimiter" must be an 1-character string
>>> for line in csv.reader(io, delimiter=';'):
...     print line
... 
['name', 'surname']
['sascha', 'gottfried']

To debug the situation dump the current value of 'self.delimiter' to the console or similar before it will be passed as a delimiter to csv.reader(). It must be a different value and/or type than ';'. Looking at the code of django-adaptors you can validate your django-adaptors model definition with this base class method as well to debug. This call should print out what you defined as Meta.delimiter.

>>> from core.models import SkuCsvModel
>>> SkuCsvModel.has_class_delimiter()

But it is pretty fine to omit a delimiter definition and call 'import_from_file' on the model. Make sure there is no class delimiter defined. If so the importer runs a CSV sniff to detect the delimiter from the file you passed. If you provide the file you mentioned, the sniffer will detect a ';' and sets self.delimiter.

>>> from core.models import SkuCsvModel
>>> SkuCsvModel.has_class_delimiter()
None
>>> my_csv_list = SkuCsvModel.import_from_file(file = open("labconco.csv"))