On 28 Apr 2013, at 10:18pm, Paolo Bolzoni wrote:
> Interesting, so sqlite3 is smart enough to actually move the blob instead
> of copying and deleting? If it is the case it is indeed great.
SQLite3 keeps all the data for a row together on disk. It rewrites the entire
row any time any field in
Interesting, so sqlite3 is smart enough to actually move the blob instead
of copying and deleting? If it is the case it is indeed great.
On Sun, Apr 28, 2013 at 5:12 PM, Simon Slavin wrote:
>
> On 28 Apr 2013, at 3:51pm, Paolo Bolzoni
> wrote:
>
>> So I should write my BLOB in another (not-inde
I use only the C API. The function causing it in my program is:
fprintf(stderr, "%s\n", sqlite3_errmsg(db));
and I think it comes out from:
sqlite3.c:70718: zErr = sqlite3MPrintf(db, "cannot open %s column
for writing", zFault);
On Sun, Apr 28, 2013 at 6:05 PM, Richard Hipp wrote:
> On Su
On Sun, Apr 28, 2013 at 10:02 AM, Paolo Bolzoni <
paolo.bolzoni.br...@gmail.com> wrote:
> I get this error: "cannot
> open indexed column for writing."
>
> What does it mean?
>
>
That error is not coming from SQLite. Are you using a wrapper program of
some kind - or perhaps a third-party query to
On 28 Apr 2013, at 3:51pm, Paolo Bolzoni wrote:
> So I should write my BLOB in another (not-indexed) table, UPDATE the
> indexed table copying from the other,
> and finally delete the line in the first table? All in one transaction?
That would work and would be a good solution if you change one
So I should write my BLOB in another (not-indexed) table, UPDATE the
indexed table copying from the other,
and finally delete the line in the first table? All in one transaction?
On Sun, Apr 28, 2013 at 4:23 PM, Simon Slavin wrote:
>
> On 28 Apr 2013, at 3:02pm, Paolo Bolzoni
> wrote:
>
>> And
On 28 Apr 2013, at 3:02pm, Paolo Bolzoni wrote:
> And it seems quite an improvement, alas now I get this error: "cannot
> open indexed column for writing."
You have a column of type BLOB.
It is now an indexed column.
You are trying to use the BLOB editing routines to write into that BLOB rather
Sorry, it seems gmail messed up the layout just before sending.
On Sun, Apr 28, 2013 at 4:02 PM, Paolo Bolzoni
wrote:
> I was playing with indexes, I started checking one my queries with
> EXPLAIN QUERY PLAN and
> I got this result:
> selectidorder fromdetail
> -- -
On 14 Jul 2011, at 5:37pm, Black, Michael (IS) wrote:
> Would glob be faster than like? Since I assume like has to case-convert?
Given your SELECT string:
>> select * from tsamov where tsamov_code like 'AFG%'
If you want speed don't use either, use
SELECT * FROM tsamov WHERE tsamov_code BETW
Would glob be faster than like? Since I assume like has to case-convert?
And you don't have to set any pragmas or NOCASE for it to use the index.
select * from tsamov where tsamov_code glob 'AFG*';
sqlite> explain query plan select * from tsamov where tsamov_code glob 'AFG*';
sele order
On 7/14/2011 12:01 PM, Pavel Ivanov wrote:
>> could the next query use the tsamov_code index ??:
>> select * from tsamov where tsamov_code like 'AFG%'
>
> Only after
> pragma case_sensitive_like = true;
... or if the index uses COLLATE NOCASE clause.
--
Igor Tandetnik
___
> could the next query use the tsamov_code index ??:
> select * from tsamov where tsamov_code like 'AFG%'
Only after
pragma case_sensitive_like = true;
Read more about it here: http://www.sqlite.org/optoverview.html#like_opt.
Pavel
On Thu, Jul 14, 2011 at 11:56 AM, Sebastian Bermudez
wro
Sorry, my emailer messed things up, try it again
0OpenVirtual10keyinfo(1,BINARY)
1Goto031
2Integer00
3OpenRead02
4SetNumColumns02
5Integer00
6OpenRead23keyinfo(1,BINARY)
7Integer1
Wilfried Mestdagh wrote:
Hi Bert,
'select distinct Name ' +
'from Rx ' +
'where RxDT >= ' + DT + ' ' +
'order by Name'
One thing is not yet clear to me. That is the 'distinct'. To have this
as fast as possible I have to make also a separate index on "RxDt,
Name". Is that correct ?
Wilfried Mestdagh wrote:
Hi,
I'm a little confused how to create indexes. (beginners question, but I
could not find or understeand clearly in FAQ). I have on a table
following 3 selects:
'select distinct Name ' +
'from Rx ' +
'where RxDT >= ' + DT + ' ' +
'order by Name'
'select * ' +
'fr
15 matches
Mail list logo
