Anne, you don't say that you make your 100 inserts a transaction.
Unless you do that the ACID feature of Sqlite will make it slow as it
has to finalize each insert onto disk before starting on the next.
Within a transaction it will do the 100 inserts then finalize the entire
100 onto disk in one write session.
When you use a transaction you also have the option of a ROLLBACK should
something go wrong before the transaction completes. Then you are not
left with incomplete data.
Like this
sqlite3_prepare
start transaction
loop
bind data values to prep'd statement
sqlite3_step
sqlite3_reset
until 100 iterations
commit transaction
sqlite3_finalize
JS
[EMAIL PROTECTED] wrote:
Hi to all
Today I inserted as a Test 100 Records in my DB. I repeat therefor
100 times the same "Insert into...". The Job starts and I wait,
and wait, and wait.
I say, Programm hangs or died, because I do a Mistake in Code.
But it isn't.
The Job needs really so much time. Trace-Message shows me, what he
just do. I look and search over a few hours in the Docu and older
Mails and search with Keywords like "Faster", "Performance",
"Disable Journal" and so on.
Some Results was successful. After I perform the "PRAGMA synchronous=OFF"
it works much faster. If I see it right, now I cannot Rollback, because
Journal is disabled. Is it so? And is it recommendable? Or is it a
"Never to this..."? ;-)
Is there perhaps any other way to make it faster (...with Customizing
some parameters)?
In the moment I'm just a little bit confused... which adjustment has
which effect?
Greetings, Anne
