python - Quickly dumping a database in memory to file -

i want take advantage of speed benefits of holding sqlite database (via sqlalchemy) in memory while go through one-time process of inserting content, , dump file, stored used later.

consider bog-standard database created in usual way:

# in-memory database e = create_engine('sqlite://') 

is there quicker way of moving contents disc, other creating brand new database , inserting each entry manually?

edit:

there doubt whether or not i'd see benefits using in-memory database. unfortunately see huge time difference of 120x.

this confusion due me missing out important detail in question. due lack of understanding on part re: caches / page sizes / etc. allow me elaborate:

i running simulations of system have set up, each simulation going through following stages:

1. make queries database.
2. make calculations / run simulation based on results of queries.
3. insert new entries database based on recent simulation.
4. make sure database date new entries running commit().

while ever make dozen or insertions on each simulation run, run millions of simulations, , results of each simulation need available future simulations take place. say, read and write process takes considerably longer when running file-backed database; it's difference between 6 hours , month.

hopefully clarifies things. can cobble simple python script outline process further little further if necessary.

sqlalchemy , sqlite know how cache , batch-inserts fine.

there no benefit in using in-memory sqlite database here, because database uses pages on-disk version would, , difference pages written disk disk-based database. difference in performance 1.5 times, see sqlite performance benchmark -- why :memory: slow...only 1.5x fast disk?

there no way move in-memory database disk-based database @ later time, short of running queries on in-memory database , executing batch inserts disk-based database on 2 separate connections.