[GRASS5] [bug #2904] (grass) v.in.sites for larger site files does not finish

[Radim Blazek]

It will be in the driver implementation of db__driver_close_database(),
either in save_table() or free_table() in grass51/db/drivers/dbf/db.c.

Can you please try to comment the free_table (i); ( row 110 )
in the grass51/db/drivers/dbf/db.c.

In theory, one driver can be reused for another database, but AKAIK
it is not in current GRASS. So it is faster just close the driver.
OTOH, it can become a problem if DBMI swith to dlopen because of windows 
port.

Radim


guest user via RT wrote:
> this bug's URL: http://intevation.de/rt/webrt?serial_num=2904
> 
> Request number 2904 was commented on by 'guest' (guest user). 
> Responding to this message will send mail to the requestor.
> 			
> 			Request Tracker
> 			rt at intevation.de
> 
> --------------------------------------------------------------
> Cc: grass5 at grass.itc.it
> 
> I tried to track down the problem - here is more detailed info, but no
> solution yet:
> 
> The transfering of sites to vector file seems to slow down as the number
> of points increases (maybe db_append_string has to go through everything
> that has been written each time it adds a new point?)
> but the real problem seems to be in db_close_database(driver) -
> the program spends by far the most time on that function,
> for illustration, 280000 points takes 3 minutes doing the transfering sites
> to vect file but it spends 8 minutes on db_close_database and it extremely
> slows down the machine (windows, mouse practically don't work on my computer
> during those 8 minutes).  Everything else runs for a negligible time.
> Slower importing of 280000 points is not really a problem, but
> for larger files (500000+) I was unable to get it past
> db_close_database(driver) after running it for hours.
> So the bug apperas to be somewhere in lib/db/dbmi_driver/d_closedb.c or
> d_close_cur.c, but I am only guessing here.
> 
> 
> 
> -------------------------------------------- Managed by Request Tracker
> 
> _______________________________________________
> grass5 mailing list
> grass5 at grass.itc.it
> http://grass.itc.it/mailman/listinfo/grass5