dse is a simple and crude way of not executing SQL queries in sequence, but caching values until a given max value has been met and then execute them using the executemany-method. The result can be huge speed gains.

dse was only tested on SQLite3 but intended for use in django as well.

import sqlite3 # for testing purposes

from dse import DelayedSqlExecutor

conn = sqlite3.connect(':memory:')
cursor = conn.cursor()
cursor.execute('create table filedata (id INTEGER PRIMARY KEY, filepath TEXT, filename TEXT, filesize INTEGER)')
d = DelayedSqlExecutor(cursor, paramtoken= '?') # using the ? paramtoken here for sqlite3. Leave it blank and it`ll use %s as support by Django etc.
d.addObject('filedata', ('id', 'filepath', 'filename', 'filesize'))

for i in range(0, 999):
    # adding some dummy data. Notice the absence of the id-field. This will trigger inserts.
    #Adding the id-field would trigger an update for data not yet in the db
    d.addItem('filedata', {'filepath': '/tmp/', 'filename': 'test%s.txt' % i, 'filesize': i})
# No SQL has been executed yet, the default limit is 1000 items

# Adding another item will trigger the execution of SQLs and reset the d-instance
d.addItem('filedata', {'filepath': '/tmp/', 'filename': 'test%s.txt' % i, 'filesize': i})

# Adding some records to update
d.addItem('filedata', {'id': 1, 'filepath': '/tmp/', 'filename': 'testmore%s.txt' % i, 'filesize': 100})

# calling close will execute any remaining SQLs
d.close()

# you might be required to call commit on the cursor to commit the data. Depends on how you set up the cursor/connection.