🚀 go-pugleaf

Michael Aresti wrote:

>My application is working with MySQL Server. Using it I can modify every 
>row in some special schema in database. Suppose that my program is running 
>and I am only viewing query response (for example "select * from 
>employees") and somebody else using MySQL Query Browser is making some 
>changes in that table (deleting few rows).
>
>My question is: is there enabled any kind of synchronization that I can 
>use in my program (if something is changing in tables in database it will 
>inform my program about it). I want that using my program I will view 
>always the most actual results.

No.  The work others do is isolated from you until you requery the 
database.  Some databases have callback APIs to inform you of record 
changes (or you can build them yourself on a table-by-table basis if you 
have to) but they're tricky to implement.  If you have a table that is 
constantly changing for which you have to see the latest changes, the best 
bet is usually just to refresh the dataset on some kind of timer.

-- 
Kevin Dean [TeamB]
Dolphin Data Development Ltd.
http://www.datadevelopment.com/

NEW WHITEPAPERS
Team Development with JBuilder and Borland Enterprise Server
Securing Borland Enterprise Server
http://www.datadevelopment.com/papers/index.html

Please see Borland's newsgroup guidelines at
http://info.borland.com/newsgroups/guide.html

Michael Aresti wrote:

> My question is: is there enabled any kind of synchronization that I can 
> use in my program (if something is changing in tables in database it 
> will inform my program about it). I want that using my program I will 
> view always the most actual results.

Kevin's answer is the only one that I know of that will work. In 
addition, each update, insert, or delete could have a change/auditing 
table associated with it with a timestamp. That way, you could use a 
threaded call to see what rows need to be updated in the refresh, if any.

Paul Nichols [TeamB] wrote:
> Kevin's answer is the only one that I know of that will work. In 
> addition, each update, insert, or delete could have a change/auditing 
> table associated with it with a timestamp. That way, you could use a 
> threaded call to see what rows need to be updated in the refresh, if any.

Thanks for help. Your both answers are very good explanation of problem 
I wanted to resolve. But I also think that at present my Java programing 
skills are insufficiently to this alone... so I decided to change the 
"Transaction Isolation" level to TRANSACTION_REPEATABLE_READ. This way 
nobody else can make any changes in database tables and I have always 
the most actual data.