A dirt simple CSV to sqlite db generator.

README.md

To test with http://www.maxmind.com/en/worldcities:

$ gzcat worldcitiespop.txt.gz | python csv2db.py worldcitiespop
$ sqlite3 worldcitiespop.db

csv2db.py

#!/usr/bin/env python
import codecs
import cStringIO
import csv
import os
import sqlite3
import sys

dbname = sys.argv[1] if len(sys.argv) > 1 else 'data'
if os.path.exists(dbname + '.db'):
    os.unlink(dbname + '.db')

dbconn = sqlite3.connect(dbname + '.db')
cursor = dbconn.cursor()

params = []
def create_table(columns):
    columns_with_types = ['%s text' % column for column in columns]
    create_stmt = 'create table %s (%s)' % (dbname, ', '.join(columns_with_types))

    global params
    params = ','.join(['?'] * len(row))
    cursor.execute(create_stmt)

def insert(row):
    param_values = [unicode(v, 'utf-8', errors='ignore') for v in row])
    cursor.execute('insert into %s values (%s)' % (dbname, params), param_values)

header = True
reader = csv.reader(codecs.EncodedFile(sys.stdin, 'utf-8', errors='ignore'))
for row in reader:
    if header:
        create_table(row)
        header = False
        continue

    insert(row)

dbconn.commit()
dbconn.close()