Right, ORDER by, oherwise the row order is undefined.
Why the urge to grab the mailing list, then look at the references?
Dunno, sloppy hobbyist..
On Sat, Apr 2, 2011 at 5:14 PM, Jay A. Kreibich wrote:
> On Sat, Apr 02, 2011 at 05:01:34PM -0700, Matt Young scratched on the
On Sat, Apr 02, 2011 at 05:01:34PM -0700, Matt Young scratched on the wall:
> sqlite> select 1 as type union select 2;
> type
> 1
> 2
> sqlite> select 'tr' as type union select 2;
> type
> 2
> tr
> sqlite>
Yes.
First, understand that the row order of all SQL queries is undefined,
unless
On Sat, Apr 2, 2011 at 8:01 PM, Matt Young wrote:
> sqlite> select 1 as type union select 2;
> type
> 1
> 2
> sqlite> select 'tr' as type union select 2;
> type
> 2
> tr
> sqlite>
>
> The order of the rows change when text replaces a numeric.
>
My understanding is that if
sqlite> select 1 as type union select 2;
type
1
2
sqlite> select 'tr' as type union select 2;
type
2
tr
sqlite>
The order of the rows change when text replaces a numeric.
___
sqlite-users mailing list
sqlite-users@sqlite.org
On Sat, Apr 2, 2011 at 11:10, Igor Tandetnik wrote:
> If possible, set up your table so that INSERT OR REPLACE works for it. E.g.
> declare "key" field unique, or primary key.
Yah, I've used INSERT OR REPLACE -- it's a nice extension that works
like a champ.
I'm writing a
Robert Poor wrote:
> On Fri, Apr 1, 2011 at 21:36, Igor Tandetnik wrote:
>> update table_a set avalue = (
>> select bvalue from table_b where akey=bkey
>> union all
>> select avalue);
>
> That also works. But at the risk of "moving the finish line during
Ian Strascina wrote:
> So is this a sqlite bug in that inconsistent results are coming from queries
> that should be functionally equivalent according to the above documentation
Realize that SQLite doesn't have a date or datetime data types. All the
calculations are
>It's a pain in the arse that one has to pay for copies of the
>ISOs. What do they think we're paying them for ?
I also find this a perverse effect totally contrary to their mission or
at least its spirit: produce good standards for public use. Without
free access to reference up-to-date
On 2 Apr 2011, at 4:30pm, Simon Slavin wrote:
> The ISO standard separates date and time with exactly 'T' or 't'.
Wrong. Sorry about that. The ISO standard specifies a capital 'T'. Other
standards which are obviously based on it suggest that a lower-case 't' is
acceptable. Apologies for
On 1 Apr 2011, at 11:31pm, Ian Strascina wrote:
> I have come across inconsistent results of queries using datetime functions
> and
> values. I have based everything I tested off of this page:
> http://www.sqlite.org/lang_datefunc.html
There's a problem with that web page. The ISO standard
I have come across inconsistent results of queries using datetime functions and
values. I have based everything I tested off of this page:
http://www.sqlite.org/lang_datefunc.html
I have a table with a /start_date/ property. The values written are UTC time
of
the format
11 matches
Mail list logo