On 10/16/2015 03:22 PM, ???_Ren? wrote:
> my code like below:
> --
>
>
> const char* sql =
> "begin;"
> "create table if not exists
I'm disappointed. I killed it after 1.5hrs, and suspected everything you
explained in the other thread. It is always true.
Que Sera Sera
dvn
On Fri, Oct 16, 2015 at 2:33 PM, Igor Tandetnik wrote:
> On 10/16/2015 3:23 PM, Don V Nielsen wrote:
>
>> limit (
>>SELECT net_non_pieces
my code like below:
--
const char* sql =
"begin;"
"create table if not exists personal("
"user_id integer primary key autoincrem
On 10/15/2015 03:35 AM, Philip Bennefall wrote:
> Hi all,
>
> Are there any plans in fts5 to add support for something equivalent to
> fts4aux? This would be of great use to me.
There is, but it's not actually finished or documented yet. It still
does a linear scan of the entire FTS index for ev
2015-10-15 18:18 GMT+02:00 Bruce Blackwell :
> I do not know if it is significant or not, but this is the first time I
> have ever seen this happen.
>
> sqlite3.dll from sqlite-dll-win32-x86-309.zip reports
> sqlite_source_id() as 2015-10-14 18:30:13
> a43bb2b5cf9c7c19edbce37c883826eb0f7ed764
>
I've not seen the RFC but you say "JSON only has 5 whitespace characters" and
then list only 4, and your patched array only has four 1's. ?Have you missed
one, or is the 5 wrong (or am I missing something)?
Graham
Sent from Samsung Mobile
Original message
From: Jan Nijtmans
Hi all,
Just noted in the jsonIsSpace[] array: According to RFC 7159, JSON
only has 5 whitespace characters, FF (0x0C) is not among them:
ws = *(
%x20 / ; Space
%x09 / ; Horizontal tab
%x0A / ; Line feed
On 10/16/2015 3:23 PM, Don V Nielsen wrote:
> limit (
>SELECT net_non_pieces
>FROM crrt_net_non [b]
>WHERE [b].zip = zip AND [b].crrt = crrt
Again, I don't think this does what you think it does. Test with at
least two rows in crrt_net_non, with different values for
On 10/16/2015 3:15 PM, Don V Nielsen wrote:
> Given the following:
>
> select * from tmp_addresses [x]
> where
>rowid in (
> select rowid
> from tmp_addresses [a]
> where a.version_id = '0060'
>and [a].zip = [x].zip and [a].crrt = [x].crrt
> limit (
>SELE
Wrapping up this thread.
What I am going to go with is the following. I'm using the initial WITH to
reduce as much as possible the number of records to be repeatedly searched:
maybe 25/30%. The addresses table currently has 4.4mm rows; it will go up
to 11.5mm when this goes to prod. I'm testing
Given the following:
select * from tmp_addresses [x]
where
rowid in (
select rowid
from tmp_addresses [a]
where a.version_id = '0060'
and [a].zip = [x].zip and [a].crrt = [x].crrt
limit (
SELECT net_non_pieces
FROM crrt_net_non [b]
WHERE [b].zip = [x].
On 10/16/15, Scott Doctor wrote:
>
> Form feeds are ... Often inserted for pagination and
> ignored by compilers or treated the same as a space character.
>
Yeah. Which is why I originally included them in the whitespace
definition for json1. But RFC7159 has other ideas. We'll conform to
RFC71
On 16 Oct 2015, at 9:22am, ???_Ren? wrote:
> int ret = sqlite3_exec(m_db, sql, 0, 0, &errmsg);
>
> under ios, it work nice.
> but under android it don't work.
What value is returned in your 'ret' variable ?
Simon.
Thanks, Dan. I'll keep my eyes on this. Thanks for the great work on FTS5!
Kind regards,
Philip Bennefall
On 10/16/2015 11:10 AM, Dan Kennedy wrote:
> On 10/15/2015 03:35 AM, Philip Bennefall wrote:
>> Hi all,
>>
>> Are there any plans in fts5 to add support for something equivalent
>> to fts4au
http://www.ecma-international.org/publications/files/ECMA-ST/ECMA-404.pdf
http://json.org/
Form feeds are allowed in strings. Form feeds are traditionally
treated the same as a space. Often inserted for pagination and
ignored by compilers or treated the same as a space character.
--
There was a problem with the original Windows DLL, and probably, when it was
compiled again, the source id was kept to the original.
-Original Message-
From: Bruce Blackwell
Sent: Thursday, October 15, 2015 12:18 PM
To: sqlite-users at mailinglists.sqlite.org
Subject: [sqlite] Version 3
The final solution at this point is Ruby. I really want to push everything
I can into Sqlite because *it is so freakin fast!* Unfortunately. I I am
just not getting it. Also unfortunately, iterating, getting data, and
updating data in scripting languages is not efficient. I've implemented a
tra
In order to compile SQLite3 without any dependency to the standard RTL for a
language without it (Delphi, in my case), I had to make the following changes
to the JSON1 extension:
- fixed unknown isalnum/isdigit function calls:
#define safe_isdigit(x) sqlite3Isdigit((unsigned char)(x))
#define
Cannot compile SQLite amalgamation with MDK-ARM (Keil). I have met errors:
|..\..\Libraries\sqlite-amalgamation-3081101\sqlite3.c(25838):error:#5:cannot
opensource input file"sys/types.h":No such fileordirectory
..\..\Libraries\sqlite-amalgamation-3081101\shell.c(59):error:#5:cannot
opensource
19 matches
Mail list logo