Marco Turco wrote:Hello Rimantes,
I'm using Sqlite and xHarbour since late 2010 in all my products and in network environments without problems.
Generally speaking, you don't need a client-server structure to operate on a network with Sqlite but you should cumulate the write calls inside a Begin Immediate->Commit sequence in order to minimize the file locktime.
To be clear, if you make
INSERT INTO TABLE 1 ....
INSERT INTO TABLE 2 ...
INSERT INTO TABLE 3 ...
you will have 3 lock-unlock loops
Instead, if you make
Begin immediata
INSERT INTO TABLE 1 ....
INSERT INTO TABLE 2 ...
INSERT INTO TABLE 3 ...
Commit
You have just a 1 lock-unlock loop
We have customers with also 25-28 client that operate without problem this way.
Are you using Sqlite3 with odbc or with a wrapper ?
Marco , you provide so PERFECT VALUABLE INFORMATION !!!
How fine that is , many thanks to you !
At first , reading info about SqLite and network problems with them I didn't begin a seriuos working ( production ) project with Sqlite . Now I have a new project , small enterpise of production , ~10-12 PC in network . So SqLite for such project is ideal - no server , minimal requirements and administration and etc.
About locking - I'm working ONLY in such manier , as you described - begin ... insert/update one or some tables ... commit . In mine main workplace I'm working with Oracle , inserting/updating I'm doing only in this way some years alreday .
Because I was afraid to begin with SqLite database in network , I begin with fireBird . But now I'll return to SqLite . With FireBird I'm working with hbfbird from harbour\contrib ( I'm working with harbour ) . I was working with hbsqlt3 from harbour\contrib . Also I'll review solution from Ssbbs .
Many thanks to you , Marco !
With best regards !
Rimantas U.