On Sep 17, 2011, at 1:06 PM, Simon Slavin wrote:
>
> On 17 Sep 2011, at 6:42pm, Petite Abeille wrote:
>
>> On Sep 17, 2011, at 7:33 PM, Simon Slavin wrote:
>>
>>> As was clear from my post, I was referring to SQL standards. What various
>>> implementation vendors choose to do is up to them. But the multi-spec
>>> syntax referred to in the OP is not in any SQL standard I've seen.
>>
>> In BNF Grammar for ISO/IEC 9075:1999 - Database Language SQL (SQL-99), under
>> contextually typed row value expression list:
>>
>> http://savage.net.au/SQL/sql-99.bnf.html#contextually%20typed%20row%20value%20expression%20list
>
> <insert columns and source> ::= <from subquery> | <from constructor> |
> <from default>
>
> ? Which one, and where is the expansion that allows for multiple sets of
> brackets after "VALUES" ?
>
All this chattering among us doesn't really matter (other than for academic
purposes). All that matters is whether or not Richard and co. deem it worth
including as a capability in factory provided SQLite. While I can't attest to
it, there are probably instances where SQLite deviates from the so called
standard ("the problem with standards is there are plenty of them" and all
that). For me, if Pg does it, the SQLite could do it if those who make SQLite
could be convinced of its usefulness.
For me, I don't care either way. I am glad Pg has it because I am trying to
convert MySQL data to Pg. While I am failing to do so painlessly for the most
part, a few of the (simpler) tables convert fine because both MySQL and Pg
support multi-line INSERTs.
Puneet.
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
