You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is there a specific reason why column names cannot start with an underscore?
I have some tables with such columns, and I cannot serve them to the community;
so, I wonder if that could be change…?
The text was updated successfully, but these errors were encountered:
On Tue, Mar 27, 2018 at 06:43:06AM -0700, Zarquan wrote:
Email from Alberto Micol at ESO
> Is there a specific reason why column names cannot start with an
> underscore? I have some tables with such columns, and I cannot
> serve them to the community; so, I wonder if that could be
> change????
The reason simply is that SQL92 doesn't allow that. I'll freely
admit that this deviation from the C identifier syntax has bitten me,
too, occasionally (though with column names).
Given the state of things, I'd guess mapping the invalid names at the
ADQL translation layer would be a reasonable way out, if renaming the
tables in question really isn't an option.
On the other hand, I'd certainly not put up much resistance to
changing the grammar of <regular identifier>. The downside is that
if there were RDMSes out there that actually implement the SQL92
grammar, it's people running them that would have to do some fairly
crazy machinations on uploaded tables (either convert regular
identifiers with underscores to delimited identifiers and then
implementing caee-insensitive resolution themselves, or use some
sufficiently robust mapping).
And they'd be punished like that through no fault of their own; one
*might* argue it's people using nonstandard table names that should
suffer...
Then again, perhaps all DBMS authors did the right thing and used the
C grammar for <regular identifier>.
Email from Alberto Micol at ESO
The text was updated successfully, but these errors were encountered: