On Mon, 27 Jan 2020 12:00:52 -0700
"Keith Medcalf" <kmedc...@dessus.com> wrote:
> If you mean importing into a pre-existing table as above that is
> declared with "check typeof(a) in ('integer', 'real')" then yes. The
> check constraint is run before affinity is applied rather than after
> (unfortunately) ...
Thank you, Keith. Yes, that's what I was referring to.
> >I imagine it's already being considered: if pragma table_info
> >included a column with the SQL for generated columns (and NULL)
> >otherwise, the shell could use that instead of SELECT * to determine
> >the number of insertable columns.
>
> That data is already in the schema structures as is the assigned
> affinity:
Good to know. I realize of course that computed columns are a new
feature. Knowing that their "computed" property is easily ascertained,
I'm optimistic that the shell's .import will one day avail itself of
that information.
What's old is always new again. ISTR when Microsoft SQL Server added
computed columns, they also got in the way of bulk-loading at first.
The rule for a while was that the buik-copy utility (bcp) couldn't be
used with such tables.
--jkl
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
