On Mon, Oct 3, 2016 at 11:51 AM, Richard Hipp wrote:
> Safe way: In a separate process, use the backup API
> (https://www.sqlite.org/backup.html) to copy the content of the main
> DB over to a separate DB, then "DELETE FROM log;" on the main DB.
> This will work without any cooperation on the par
Hello Richard !
Ok I missed this point, but still while fighting to use sqlite3 with big
databases I was thinking on some custom changes to allow sqlite relax
some restrictions:
1- Create a new sqlite reserved table for register attached databases,
this way every time a program try to open an
On 10/3/16, Domingo Alvarez Duarte wrote:
> Hello !
>
> Thinking about this and the problem I'm experiencing with big databases
> with sqlite3 "vacuum" probably could be a good idea to use a flag in the
> sqlite3 header to inform other processes to reopen the database.
>
That might work on unix.
Hello !
Thinking about this and the problem I'm experiencing with big databases
with sqlite3 "vacuum" probably could be a good idea to use a flag in the
sqlite3 header to inform other processes to reopen the database.
Right now every time sqlite3 would perform an operation on a database it
a
It seems that you just want to keep access to all of your historic
logging so rather than copy/backup the entire database you could just
create a new archive DB (or open an old one), attach it, copy x
records to the archive and then delete the same x records from the
master.
How big is your log da
Are you looking to keep the logs in the same file, or, are you looking to
put your log entries in a new file?
If you're interested in just keeping a single file, and if you've got
access to change your code that is writing to the database, then, what I
would do is "create table if not exists Logs
* Richard Hipp:
> You cannot rename a database file while another process has that
> database open. Windows simply will not allow that. If you do it on
> Unix, then the process that has the file open will not know that the
> file has been renamed and will continue to write to the original file.
> On Oct 3, 2016, at 11:18 AM, Luca Ferrari wrote:
>
> What is the right way to do it without having to stop the application
> (and therefore without knowing when a new I/O operation will be
> issued)?
You could use the attach/detach [1] functionality to transparently roll the
logs over.
Some
On 10/3/16, Luca Ferrari wrote:
> Hi all,
> in one of my application I use a sqlite3 database as a log of
> activity. As you can imagine the file grows as time goes by, so I'm
> figuring I've to substitute it with an empty one once a good size is
> reached.
> What is the right way to do it without
On 3 Oct 2016, at 10:18, Luca Ferrari wrote:
> in one of my application I use a sqlite3 database as a log of
> activity. As you can imagine the file grows as time goes by, so I'm
> figuring I've to substitute it with an empty one once a good size is
> reached.
> What is the right way to do it with
Hi all,
in one of my application I use a sqlite3 database as a log of
activity. As you can imagine the file grows as time goes by, so I'm
figuring I've to substitute it with an empty one once a good size is
reached.
What is the right way to do it without having to stop the application
(and therefor
11 matches
Mail list logo
