Django（或sqlite3）更改模型属性的顺序

Question

I'm using django with sqlite3, and here is my models.py :

from django.db import models
from django.conf import settings
from django.utils import timezone

class VoterDetails(models.Model):
    number = models.CharField(max_length=200,null=True)
    voter_id = models.CharField(null=True,max_length=200)
    elector_name = models.CharField(max_length=200,null=True)
    father_or_husband_name = models.CharField(max_length=200,null=True)
    has_husband = models.CharField(max_length=200,null=True)
    house_no = models.CharField(max_length=200,null=True)
    age = models.CharField(max_length=200,null=True)
    sex = models.CharField(max_length=6,null=True)
    ac_name = models.CharField(max_length=200,null=True)
    parl_constituency = models.CharField(max_length=200,null=True)
    part_no = models.CharField(max_length=200,null=True)
    year = models.CharField(max_length=200,null=True)
    state = models.CharField(max_length=200,null=True)
    filename = models.CharField(max_length=200,null=True)
    main_town = models.CharField(max_length=200,null=True)
    police_station = models.CharField(max_length=200,null=True)
    mandal = models.CharField(max_length=200,null=True)
    revenue_division = models.CharField(max_length=200,null=True)   
    district = models.CharField(max_length=200,null=True)
    pin_code = models.CharField(max_length=200,null=True) 
    polling_station_name = models.CharField(max_length=200,null=True)
    polling_station_address = models.CharField(max_length=200,null=True)
    net_electors_male = models.CharField(max_length=200,null=True)
    net_electors_female = models.CharField(max_length=200,null=True) 
    net_electors_third_gender = models.CharField(max_length=200,null=True)
    net_electors_total = models.CharField(max_length=200,null=True)
    change = models.CharField(default="",max_length=200,null=True)

    def __str__(self):
        return self.elector_name

As one can see, there is no attribute, which would claim to be the primary key . In this case, django generates an attribute id , which is an auto incrementing Auto field , which is chosen as the primary key.

I did the migrations, and then checked the schema of the table temp_voterdetails thus generated.( temp is the name of my app.)

sqlite> .schema temp_voterdetails
CREATE TABLE IF NOT EXISTS "temp_voterdetails" (
"id" integer NOT NULL PRIMARY KEY AUTOINCREMENT,
"age" varchar(200) NULL,
"ac_name" varchar(200) NULL,
"change" varchar(200) NULL,
"district" varchar(200) NULL, 
"elector_name" varchar(200) NULL, 
"father_or_husband_name" varchar(200) NULL, 
"filename" varchar(200) NULL, 
"has_husband" varchar(200) NULL, 
"main_town" varchar(200) NULL, 
"mandal" varchar(200) NULL, 
"net_electors_female" varchar(200) NULL, 
"net_electors_male" varchar(200) NULL, 
"net_electors_third_gender" varchar(200) NULL, 
"net_electors_total" varchar(200) NULL, 
"number" varchar(200) NULL, 
"pin_code" varchar(200) NULL, 
"police_station" varchar(200) NULL, 
"polling_station_address" varchar(200) NULL, 
"polling_station_name" varchar(200) NULL, 
"revenue_division" varchar(200) NULL, 
"sex" varchar(6) NULL, 
"state" varchar(200) NULL, 
"voter_id" varchar(200) NULL, 
"year" varchar(200) NULL, 
"house_no" varchar(200) NULL, 
"parl_constituency" varchar(200) NULL, 
"part_no" varchar(200) NULL);

As you can see, the order of attributes is alphabetical, and thus is changed from the original order of declaration of the attributes in the models.py file.

This thing aside, I tried to import a CSV (that followed the structure defined in the models.py file) by:

.mode csv
.separator ","
.import /path/to/my/csv temp_voterdetails 

Then I get the error (times the number of rows, obviously):

INSERT failed: UNIQUE constraint failed: temp_voterdetails.id
/path/to/my/csv:164705: expected 28 columns but found 27 - filling the rest with NULL

So, initially I thought that I'm not giving any input for the "id", so it must be auto generated, and thus should never fail the UNIQUE constraint. But alas to the alphabetic shuffling of django to sqlite3 conversion, id seems to input from number , and at the end there is a single column left. Because of which the UNIQUE constraint fails, and the last column is left unused.

How do I make sure that the schema of the table I create is correctly designed and is same as the model defined in models.py ?

---

UPDATE: After the automatic alphabetical ordering of attributes is stopped, I'll add an attribute global_number (AUTO INCREMENT, primary key) just at the end. All the other attributes might be matched with the 27 existing column headers, and the last one, since isn't present in the CSV file, will be generated automatically, thus generating my primary key.

This is what I plan. Correct me if this is wrong in any way.

---

Answer 1

The attributes seem to always be ordered alphabetically, in the sqlite3 model, because of which we can't directly import our csv using sqlite3. The main problem I faced was the slow speed of row addition to the table. This was because I was doing 1 query per row addition. This can be easily solved by making a list of entry objects and bulk adding this to the table. This takes only a single query, thus takes considerably less time. Here is the code:

import os
import django
os.environ['DJANGO_SETTINGS_MODULE']='<your-site>.settings'
django.setup()
from temp.models import <your-model>
import csv
with open('/path/to/csvfile','r') as f:
    reader = csv.reader(f)
    lst = []
    for i,row in enumerate(reader):
        print('Adding :'+str(i))
        if(i!=0):
            lst.append(<your-model>(
                name = row[0], #attribute 1
                popularity = row[1], #attribute 2
                ssum = row[2], # attribute 3
                yearrange = row[3], #attribute 4
                state = row[4] # attribute 5
                ))
    <your-model>.objects.bulk_create(lst)