On Thu, Sep 26, 2019 at 7:13 PM Alvaro Herrera <
alvherre@2ndquadrant.com> wrote:
>
> On 2019-Sep-25, Asim R P wrote:
>
> > I reviewed your patch today. It looks good overall. My concern is that
> > the ExecFetchSlotHeapTuple call does not seem appropriate. In a generic
> > place such as createas.c, we should be using generic tableam API only.
> > However, I can also see that there is no better alternative. We need to
> > compute the size of accumulated tuples so far, in order to decide whether
> > to stop accumulating tuples. There is no convenient way to obtain the
> > length of the tuple, given a slot. How about making that decision solely
> > based on number of tuples, so that we can avoid ExecFetchSlotHeapTuple call
> > altogether?
>
> ... maybe we should add a new operation to slots, that returns the
> (approximate?) size of a tuple? That would make this easy. (I'm not
> sure however what to do about TOAST considerations -- is it size in
> memory that we're worried about?)
That will help. For slots containing heap tuples, heap_compute_data_size() is what we need. Approximate size is better than nothing.
In case of CTAS, we are dealing with slots returned by a scan node. Wouldn't TOAST datums be already expanded in those slots?