Does this risk implies a database corruption or the risk only implies the undesired behaviour, from the logical point of view, of having two versions of the same stored procedure in two different transactions ( situation that ends when the older transaction commits or rolls back ) ?

Aldo Caruso

 

23.03.2014 04:17, Aldo Caruso wrote:
>
> Frequently I modify stored procedures and triggers.
>
> If some user were connected before I started my transaction (the
> one which changes the stored procedure) he will still be using the
> unmodified version of the stored procedure. On the other hand, users who
> connects after I commit my transaction will be using the modified version.

True.

> In that case is there a risk that two different versions of the
> same stored procedure coexist (in different transactions)

Surely this may be undesired, but this "risk" is a documented behavior,
AFAIK.

Dmitry