For the last point, using the SQLITE_OMIT_TEMPDB option, did you compile from
the amalgamation or the full cannonical sources? According to
https://www.sqlite.org/compile.html#_options_to_omit_features, many "OMIT"
options only work when the cannonical sources are used (my guess if compiling
From https://www.sqlite.org/c3ref/c_alter_table.html
"The 5th parameter to the authorizer callback is the name of the database
("main", "temp", etc.) if applicable."
Is that not the case in your tests?
I guess the difference between "CREATE TEMP TABLE x" and "CREATE TABLE temp.x"
ist hat the
Hello,
I found some behaviors that should probably be considered bugs (and should be
fixed and/or documented).
Let’s start the explanation by observing some behavior that actually is correct
and consistent. Below, we observe which type of action is reported by
sqlite3_set_authorizer(), given
On 3 Feb 2019, at 9:52am, Gerlando Falauto wrote:
> I do want them sorted, and I also want the whole (huge) dataset to be
> processable without having to store it all in memory or temp files.
> Sounds like the whole purpose of an index, doesn't it?
> I do know SQL is all about the result, not
Simon, Tim,
Il sab 2 feb 2019, 23:40 Simon Slavin ha scritto:
> On 2 Feb 2019, at 10:19pm, Gerlando Falauto
> wrote:
>
> > Results should be naturally ordered by source1, source2,ts.
>
> [Sorry, I missed this the first time. Thanks, Tim.]
>
> Sorry, no. You're making assumptions about how
5 matches
Mail list logo
