There is no datetime type in SQLite. You are storing ISO Text representations
so you should declare the column as TEXT.
An Index is only useable for a prefix of equality contstraints followed by ONE
inequality constraint.
From your index (model_id, confidence, ts) and your query constraints (
This is the expected and documented behaviour. Maybe you are looking for UPSERT?
-Ursprüngliche Nachricht-
Von: sqlite-users [mailto:sqlite-users-boun...@mailinglists.sqlite.org] Im
Auftrag von Roman Fleysher
Gesendet: Dienstag, 23. Oktober 2018 19:53
An: General Discussion of SQLite Data
Hi,
I have built versions of the commandline shell and the dll so as to get
access to the geopoly module. My builds were done firstly using the
mingW64 compiler (from the command prompt) on windows10, and then later
using codeblocks. For both builds I specified the same set of compile
time direct
>
> I found code of conduct in documentation and I was wondering if it were
> true. Checking the version history it appears to have been added on
> 2018-02-22.
>
Sure that publishing date wasn't 2018-04-01?
On Wed, 24 Oct 2018 at 08:02, Stefan Evert wrote:
>
> > On 23 Oct 2018, at 07:04, Paul
EXPLAIN QUERY PLAN shows the "High Level" outline of the plan for executing
your query, primarily the constraints imposed on indexes, but not the WHERE
conditions that are not used to constrain an index lookup.
It does not show the "code" that is executed. Use EXPLAIN rather than EXPLAIN
QU
On 23 Oct 2018, at 11:45pm, Hamesh Shah wrote:
> CREATE INDEX `detected_model_id_confidence_ts` ON `detected` (
> `model_id`,
> `confidence` ASC,
> `ts` ASC
> );
Create another index with the fields in this order
> `model_id` ASC,
> `ts` ASC,
> `confidence` ASC
and try it again. By the way,
I’m not the expert here, but it appears that the cause is that your looking for
things greater than some confidence. This forces an index scan. There’s nothing
that gives a list of different confidences greater than, in this case .8, but
even if it did, an index scan might be faster than individ
I need a little help with some strange indexing behaviour.
I have a table called detected.
i create a index for:
id integer, confidence ASC, timestamp ASC
Then when I query with a simple select from where with integer, then
confidence, then timestamp in order, for some reason the timestamp index
> On 23 Oct 2018, at 07:04, Paul wrote:
>
> If my opinion has any value, even though being atheist, I prefer this CoC 100
> times over
> the CoC that is being currently pushed onto the many open-source communities,
> that was
> created by some purple-headed feminist with political motives. Thi
> Le 23 oct. 2018 à 19:52, Roman Fleysher a
> écrit :
>
> I am using INSERT OR REPLACE ... it seems that "replace" is implemented as
> "delete then insert" rather than "update". Is that normal and expected? Am I
> doing something wrong?
Normal and expected. Check https://www.sqlite.org/lang_
That's normal. It deletes the conflicting row and inserts a new one.
On Tue, Oct 23, 2018, 10:58 AM Roman Fleysher <
roman.fleys...@einstein.yu.edu> wrote:
> Dear SQLiters,
>
> I am using INSERT OR REPLACE to update a table which holds a column which
> servers as a foreign key. But I noticed a st
Dear SQLiters,
I am using INSERT OR REPLACE to update a table which holds a column which
servers as a foreign key. But I noticed a strange behavior: If the parent
record existed, then replace mechanism replaces it, but the records from
children tables are deleted. The foreign key is set up to c
And did you confirm the confirmation requests? If you unsubscribe
without using your password, then mailman will send a confirmation
request that you need to use to actually confirm that you want to
unsubscribe (otherwise anyone could unsubscribe anyone). Clicking on the
link is generally more reli
Because it knows you don't really want to leave.
-Original Message-
From: sqlite-users On Behalf Of
Rob Dixon
Sent: Tuesday, October 23, 2018 12:41 PM
To: SQLite mailing list
Subject: Re: [sqlite] unsubscribe
I did that yesterday 3 times, got 3 confirmations and yet..
On Tue, Oct 23,
I did that yesterday 3 times, got 3 confirmations and yet..
On Tue, Oct 23, 2018 at 10:38 AM Tim Streater wrote:
> On 22 Oct 2018, at 20:08, thomgrayr...@printeasy.net wrote:
>
> > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
>
> Go to that web site, which is shown in eve
On 22 Oct 2018, at 20:08, thomgrayr...@printeasy.net wrote:
> http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
Go to that web site, which is shown in every mail you have received from this
list.
--
Cheers -- Tim
___
sqlite-u
On 10/23/2018 03:13 AM, Peter Ďurica wrote:
Table with sample data:
*create table t(a int, b int);*
*insert into t values(1,11);*
*insert into t values(2,12);*
now query using any window function (row_number, rank, ) after UNION or
UNION ALL will cause sqlite.exe crash (no regular error)
fo
On Monday, 22 October, 2018 14:13, Peter Ďurica wrote:
>Table with sample data:
>*create table t(a int, b int);*
>*insert into t values(1,11);*
>*insert into t values(2,12);*
...
What is up with the asterisks, they make copying VERY VERY VERY VERY difficult.
If you want to put "stars" around
I dug a little more into this with a debug build; was able to get the
same crash trace with the slightly smaller query
CREATE TABLE t(a);
SELECT 1, 1
UNION ALL
SELECT a, RANK() OVER (ORDER BY a) FROM t;
which fails the pTab!=0 assertion in sqlite3ColumnsFromExprList.
It seems lik
Table with sample data:
*create table t(a int, b int);*
*insert into t values(1,11);*
*insert into t values(2,12);*
now query using any window function (row_number, rank, ) after UNION or
UNION ALL will cause sqlite.exe crash (no regular error)
for example:
*select a, rank() over(order by b)
20 matches
Mail list logo