Re: Modifying a table while using next()

[Daniel Stephens <daniel@...>]

Vaughan McAlley wrote:
> ... the behaviour is undefined. Does this just mean it may or may not 
> include t.d during the current traversal? I can live with that. Or 
> would it produce an error?
As the behaviour is undefined, any of the possibilities may occur (or 
for that matter, something entirely different such as re-traversal of 
existing entries).

Have you considered a 2 table solution where you iterate over one table 
while writing updates into a second?