Django - cannot get loaddata to work - sql

Question

I have been using a Django database for a while without any major issues. Today I needed to set default values for a few tables for the first time. I created a fixtures directory in the top-level Django directory. Then I created the files for the default values. However, I keep getting error messages and I'm not sure why.
First I tried to use .sql files. It is worth noting that these tables are very simple; they only have one value, "name". My SQL file looked like this:
INSERT INTO MyTable (name) VALUES ('Default');
I saved this as MyTable.sql. When I ran the command python manage.py loaddata fixtures/MyTable.sql, I got this error message:
CommandError: Problem installing fixture 'MyTable': sql is not a known serialization format.
(Note: I also tried without the fixtures/ part, for the above example and the next, and got identical results).
I asked my project lead and he said he doesn't think SQL files can be used for this. So, I tried JSON files. My MyTable.json looked like this:
[
{
"model": "mydatabase.MyTable",
"pk": 1,
"fields": {
"name": "Default"
}
}
]
I'll be very upfront to admit I've never worked with JSON in this context before, only in web development, so I don't know if the issue may be something I'm doing wrong here. I tried to base it on the formatting I found here. When I ran this through the loaddata function again, I got this error message:
C:\Python27\lib\site-packages\django-1.6.1-py2.7.egg\django\core\management\commands\loaddata.py:216: UserWarning: No fixture named 'fixtures/MyTable' found.
This is my first time doing this and I've had a bit of a hard time finding documentation to figure out what I'm doing wrong. Could anyone please offer advice? Thanks!

Answer 1

As for me, the matter is the suffix of the output file.
python manage.py dumpdata -o my_dump
python manage.py loaddata < my_dump # it fails
# change the file name my_dump to my_dump.json
python manage.py loaddata < my_dump.json # it works
So, I guess the dumpdata implicitly use json as output format. While loaddata needs a format tip from file name suffix.

Answer 2

We can do this like below
models.py
from django import models
class Model_Name(models.Model):
name = models.CharField(max_length=50)
def __str__(self):
return self.name
fixures
[
{
"model": "app_name.model_name",
"pk": 1,
"fields": {
"name": "My app name"
}
}
]
command
./manage.py loaddata app_name/fixtures/model_name.json
Output: Installed 1 object(s) from 1 fixture(s)

Answer 3

To create your fixture files start with an empty database and then add some data to your database using the djanog-admin or the django shell or even pure SQL. After that you can do a
python manage.py dumpdata # to dump all your data, or
python manage.py dumpdata app_name # to dump all data of a specific app, or
python manage.py dumpdata app_name.model_name # to dump all data of a specific model
The above will print data to your stdout. In order to write it to a file use a redirect (>), for instance
python manage.py dumpdata auth.User > user_fixture.json
Update: I just saw that you are using Windows -- remember to load your fixtures using a backslash (\).

Answer 4

For those who still having this issues.
The error is telling you that:
CommandError: Problem installing fixture 'MyTable': sql is not a known serialization format.
You could look on it as:
sql is not a known serialization format.
It will not accept any file not ending on .json.
So to solve your problem use this command:
$ python manage.py loaddata <your_dump_filename>.json