2007/12/4 Ingo Godau-Gellert <[EMAIL PROTECTED]>:
> What is really strange is that FTS3 search phrases like
> SELECT referenzcode FROM volltext where volltext match ('installation
> manual') are performed really fast within some milliseconds, independent
> to the search phrase.
> But in general I a
Sorry for the delayed response. Was waiting for time to dig into this a little.
Nov 18, 2007 2:05 AM Wang Yun <[EMAIL PROTECTED]>:
> I insert rfc txt files into a full text search table, 4119 txt files are
> 188MB totally. After insert, database file is 443MB.
> Logic is below, it's not the real
Also http://code.google.com/apis/gears/api_database.html#sqlite_fts
might add some information.
I have an item on my todo list which goes something like "Refactor the
FTS docs on the wiki". Unfortunately, it's been on my todo list for a
couple months, now, with little progress.
-scott
On Nov 2
On Nov 18, 2007 3:28 PM, P Kishor <[EMAIL PROTECTED]> wrote:
> Following those pretty much to the t, I almost got everything working.
> Except, I got the following during make
>
> ../sqlite-3.5.2/ext/fts1/fts1.c:7:2: error: #error fts1 has a design
> flaw and has been deprecated.
> make: *** [fts1.
On Nov 15, 2007 8:20 AM, Dennis Cote <[EMAIL PROTECTED]> wrote:
> When I saw the Developers link I thought it was a link to the bios and
> pictures etc. for you and the other developers that had been requested
> as an addition to the new website. I never even clicked it to see what
> was there (whi
te:
> There seems to be no simple solution to that problem of fonts not
> scaling precisely other than to use images for the captions. If someone
> has one, let us know.
>
>
> Scott Hess wrote:
> > I notice that in Firefox on Linux with a maximized window on a
> > 16
Note that mod-11 will give you 11 items, 0..10!
Squid (and Apache's mod_proxy) use a system something like:
- hash the URL.
- hex-encode or base64-encode the hash.
- peel off 2 or 3 characters at a time for as many directories deep
as you want.
- the remaining characters are the filename in t
I notice that in Firefox on Linux with a maximized window on a
1600x1200 screen, the "Support" link in the navbar wraps around. The
navbar still looks nice, but since it's only half the width of my
screen, it shouldn't need to wrap. It also happens with narrower
browser windows, I'm guessing most
On Oct 30, 2007 7:18 AM, <[EMAIL PROTECTED]> wrote:
> This is still just an idea. If you think that adding a new
> required sqlite3_initialize() interface would cause serious
> hardship for your use of SQLite, please speak up now.
I think this would cause some hardship for dynamically-loaded
lib
On 10/17/07, Uma Krishnan <[EMAIL PROTECTED]> wrote:
> Yes. Makes sense (not to cache query results for embedded apps).
> So what is cached. Just dirty pages? or are raw tables cached when
> queried?
SQLite implements a tables and indices as btrees over a pager layer.
The pager layer caches pages.
On 10/17/07, Trevor Talbot <[EMAIL PROTECTED]> wrote:
> On 10/17/07, Uma Krishnan <[EMAIL PROTECTED]> wrote:
> > One other question, when a query is issued, does SQLite cache the results,
> > so that future queries can be processed off the cache (I think not)
>
> Like the "query cache" in some oth
Could it be that you're seeing the btree optimization around in-order
insertion? From btree.c:
#ifndef SQLITE_OMIT_QUICKBALANCE
/*
** A special case: If a new entry has just been inserted into a
** table (that is, a btree with integer keys and all data at the leaves)
** and the new entry
Your trigger can record the change notification in a separate table.
So long as the system processing the change notifications cleans them
up as it goes, this can be reasonably efficient.
-scott
On 10/13/07, Vladimir Stokic <[EMAIL PROTECTED]> wrote:
>
> Spot on! I could monitor the change of th
Triggers won't help, either, because triggers run in the sqlite3
handle which, um, triggers. So if app A makes a change, the trigger
runs in app A, but not in app B.
You could have a custom function which the trigger invokes to send an
IPC over to app B.
-scott
On 10/12/07, Vladimir Stokic <[E
The attached patch seems to do it. The thinking-out-loud patch in my
earlier email wasn't right (I'd kept an interim edit from the
PRAGMA-based approach).
I can't think of a reason to have EXCLUSIVE as the default.
-scott
On 10/10/07, Scott Hess <[EMAIL PROTECTED]>
On 10/10/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> "Scott Hess" <[EMAIL PROTECTED]> wrote:
> > We've just had a bit of discussion on the Google Gears team about some
> > cases where failure of an UPDATE/DELETE/INSERT while within a
> > tran
.
>
> A pragma that could configure the default begin
> "deffered/immediate/exclusive" would be nice :)
>
> Ken
>
>
> Scott Hess <[EMAIL PROTECTED]> wrote: To clarify, this is for Google Gears, a
> JavaScript library which
> includes a Database compon
Clarify^2: I'm suggesting for our use of SQLite in Google Gears. NOT
for SQLite itself. Though Ken's suggestion of a PRAGMA might be
interesting for SQLite core...
-scott
On 10/10/07, Scott Hess <[EMAIL PROTECTED]> wrote:
> To clarify, this is for Google Gears, a JavaS
n the appropriate places. This is a little bit of a
departure from using SQLite in an embedded environment.
-scott
On 10/10/07, John Stanton <[EMAIL PROTECTED]> wrote:
> If you are going to use BEGIN IMMEDIATE why not just enclose the
> transaction in some form of lock like a mute
We've just had a bit of discussion on the Google Gears team about some
cases where failure of an UPDATE/DELETE/INSERT while within a
transaction is unexpected. Well, that and that when you're
multi-threaded you can hit some hard-to-understand cases.
One suggestion was to use BEGIN IMMEDIATE for e
7, Andy Goth <[EMAIL PROTECTED]> wrote:
> On Fri, 5 Oct 2007 09:41:27 -0700, Scott Hess wrote
> > On 10/5/07, Andy Goth <[EMAIL PROTECTED]> wrote:
> > > proc sql_expand {varname} {
> > >upvar 1 $varname var
> > >set result [list]
> > >
You really should be using an SQLite-specific quote function
somewhere. But ... I don't see one in there (I'd have expected it to
be something like [db quote $arg]). You could work around it by doing
something like [db eval {select quote($arg)}], but that feels clunky.
The quoting you're using w
On 9/19/07, Ralf Junker <[EMAIL PROTECTED]> wrote:
> >Regarding per-row versus per-column tokenizers, your suggesting to
> >have something like 'content_kr TOKENIZER icu_kr' for each variant is
> >reasonable, but I'm not certain what gain it would have over simply
> >having separate tables for each
but that seems like a project for some future quarter. I
think doing anything automagic with query tokenization is a challenge,
because queries tend much more towards sentence fragments.
-scott
On 9/18/07, Ralf Junker <[EMAIL PROTECTED]> wrote:
> Hello Scott Hess,
>
> >In the in
t;DEFAULT" (or other suggestive word TBD), you're
indicating that you want the ability to customize the tokenizer per
row.
-scott
On 9/17/07, Scott Hess <[EMAIL PROTECTED]> wrote:
> As part of doing internationalization work on Gears, it has been
> determined that it is unlikely tha
As part of doing internationalization work on Gears, it has been
determined that it is unlikely that you can just define a global
tokenizer that will work for everything. Instead, in some cases you
may need to use a specific tokenizer, based on the content being
tokenized, or the source of the con
To help isolate if it's actually a library path problem, you might try
using the full path. So, instead of:
select load_extension('fts2');
do:
select load_extension('/path/to/libfts2.so');
Modified for Windows as appropriate (sorry, I mostly use Linux and
static linking). If that works,
[Just some background, while I'm thinking about it.]
Google has placeholder syntax, if you search for 'full * search', the
first hit is a Wikipedia article on 'full text search'. It's not the
same, but it is in a similar ballpark. For instance, you might have
'full */10 search' (Google doesn't,
After being out of the office for a bit, I came back and thought "My
to-do list is too long." So instead of working on the hardest bit, I
decided to see how many pieces of low-hanging fruit I could knock
off...
So today I'm getting rid of all these to-do items around fts1/2 rowid
breakage and fts
In general, you have to do something like:
- for newly-created files, create them during the transaction. If
the transaction rolls back, delete the file. If the transaction
commits, leave them alone.
- for deleted files, leave them in place. If the transaction
commits, delete the files. If
You could invert your solution. Create a temporary table which
contains all of the existing keys, and every time you insert a new
item, delete that item's key from the temporary table. At the end, do
something like 'DELETE FROM main_table WHERE key IN (SELECT key FROM
tmp_table)'.
-scott
On 9/6
On 9/3/07, Babu, Lokesh <[EMAIL PROTECTED]> wrote:
> One more question, As you said, Full text builds an index of data, so
> I hope you have done some memory analysis too, could you please tell
> me the memory usage based on your analysis.
That's really something that you'll wan to do with your sp
On 9/2/07, Babu, Lokesh <[EMAIL PROTECTED]> wrote:
> Does anyone help me out, What is FTS1, How to use FTS1, If any sample
> programs to use FTS1 and understand better.
The "fts" modules are "fulltext search" modules for SQLite. "Fulltext
search" meaning that it builds an index based on terms in
Unfortunately, the reason fts2 couldn't be "fixed" was because you
can't perform the necessary ALTER TABLE if the column you're adding is
a primary key. Since the only alternative would be to build a new
table and copy everything over, it seemed more reasonable to just let
the app developer do tha
ext/fts3.c in the current code fixes the fts2-vs-vacuum problem by
adding "docid INTEGER PRIMARY KEY" to the %_content table. This
becomes an alias for rowid, and thus causes vacuum to not renumber
rowids. It is safe to add that column because the other columns in
%_content are constructed such t
In either case, if you use the UUID as a primary key in more than one
table, you should consider having one table to convert the UUID to a
64-bit id, and use that as the primary keys on the other tables. If
you have UUID as a primary key, your table will have 2 b-trees, one
for the index of UUID t
a suspicion that this kind of index requires a
materially different model than fts has been using, which might
encourage it to be a completely different virtual table.
-scott
On 8/29/07, Scott Hess <[EMAIL PROTECTED]> wrote:
> A primary constraint of the porter algorithm in fts is
fts2 experience, it will probably need 2 or 3
weeks beyond that to really settle into a usable state.
-scott
On 8/29/07, brian kruse <[EMAIL PROTECTED]> wrote:
> On 8/24/07, Scott Hess <[EMAIL PROTECTED]> wrote:
> >
> > My current focus for the next generation is in
conjunction
> to snowball/potter). Which would be a good n-gram algorithm to implement.
>
> Finally, what's the rationale in having sqlite's own search. Why not use
> something like luceneC?
>
> Thanks in advance
>
> Uma
>
> Scott Hess <[EMAIL PROTECTED]&g
advance
>
> "Cesar D. Rodas" <[EMAIL PROTECTED]> wrote:
> On 23/08/07, Scott Hess wrote:
> > On 8/20/07, Cesar D. Rodas wrote:
> > > As I know ( I can be wrong ) SQLite Full Text Search is only match with
> > > hole
> > > words right
focus.
-scott
On 8/23/07, Cesar D. Rodas <[EMAIL PROTECTED]> wrote:
> On 23/08/07, Scott Hess <[EMAIL PROTECTED]> wrote:
> > On 8/20/07, Cesar D. Rodas <[EMAIL PROTECTED]> wrote:
> > > As I know ( I can be wrong ) SQLite Full Text Search is only match with
&
On 8/20/07, Cesar D. Rodas <[EMAIL PROTECTED]> wrote:
> As I know ( I can be wrong ) SQLite Full Text Search is only match with hole
> words right? It could not be
> And also no FT extension to db ( as far I know) is miss spell tolerant,
Yes, fts is matching exactly. There is some primitive suppo
If you're already willing to use a ram drive, you should benchmark
SQLite against a disk-backed file, with 'PRAGMA synchronous = off;'.
For most modern operating systems, this shouldn't provide
significantly different performance than using a RAM disk.
Note that you said 'Data persistence is not r
fts1 and fts2 have a design flaw which assumes semantics for the
standard SQLite rowid which aren't actually provided by SQLite across
calls to VACUUM. fts3 will fix this, either making the rowid be a
persistent idenfier across VACUUM, or by exposing a new id (or docid)
column which operates as a
At this time I do not plan to provide a way to distinguish "INTEGER
PRIMARY KEY AUTOINCREMENT" and "INTEGER PRIMARY KEY" for fts3. fts is
in the business of doing a fulltext index, not enforcing constraints!
If you require that level of operation, the appropriate solution would
be to have a parall
+1 for fts3 or fts2_1 :-)
>
> ---
> We're Hiring! Seeking a passionate developer to join our team building
> products. Position is in the Washington D.C. metro area. If interested
> contact [EMAIL PROTECTED]
>
> -Original Message-
> From: Scott Hess [
On 8/15/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
> If you find a way to get sqlite3 to re-parse the schema after your direct
> sqlite_master change, please post it to the list. I don't think it can
> be done without modifying the code or making a new connection.
You could probably manage it by do
On 8/14/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> "Scott Hess" <[EMAIL PROTECTED]> wrote:
> > I was getting ready to checkin the rowid-versus-fts2 fix, and wanted
> > to add one last bit, to upgrade older tables.
> >
> > Unfortunately, co
I was getting ready to checkin the rowid-versus-fts2 fix, and wanted
to add one last bit, to upgrade older tables.
Unfortunately, code of the form:
ALTER TABLE x_segments ADD id INTEGER PRIMARY KEY;
is documented as not supported.
http://www.sqlite.org/lang_altertable.html . As far as I can
On 7/27/07, John Stanton <[EMAIL PROTECTED]> wrote:
> Scott Hess wrote:
> > On 7/26/07, Richard Klein <[EMAIL PROTECTED]> wrote:
> >>According to the Mozilla article referenced above, it's even worse than
> >>that: *All* cache pages, dirty or not, are
On 7/26/07, Richard Klein <[EMAIL PROTECTED]> wrote:
> According to the Mozilla article referenced above, it's even worse than
> that: *All* cache pages, dirty or not, are freed at the end of *every*
> transaction, even if the transaction consisted of only read operations.
I believe this is no lo
ncerned with a good fix than a
fast fix. That and I have other demands on my time. Probably will be
fixed this coming week.
-scott
On 7/21/07, Adam Megacz <[EMAIL PROTECTED]> wrote:
"Scott Hess" <[EMAIL PROTECTED]> writes:
> In fts tables all columns other than rowid are
[Note that you may need to reload or shift-reload the page. I checked
three times across the day for the new content, and was just about to
send drh a notice that it wasn't there, when I tried a shift-reload!]
On 7/20/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
SQLite Version 3.4.1 is now
In past non-SQLite contexts, I've found that the following usually
works very well, without needing snapshot capability:
1. copy the db file.
2. lock the db.
3. copy the db file.
4. unlock the db.
The first copy hopefully pulls the entire file into memory buffers,
making the next copy very effic
Long ago and far away, I build a database abstraction layer which used
?@ for this. So you'd say something like:
stmt = prepare("select * from table where xyz in (?@)");
bind_array(stmt, 0, arrayRef);
The library would take the array, quote each element, and separate
them with commas. It w
on is to add code to fts2 and fts1 to upgrade tables, and put
a prominent disclaimer somewhere.
-scott
On 7/17/07, Scott Hess <[EMAIL PROTECTED]> wrote:
[Forwarding gist of an offline conversation with Joe.]
Looks about like what my patch looks like. Needs to additionally
handle %_seg
me something.
Tell me what you think.
--- Scott Hess <[EMAIL PROTECTED]> wrote:
> I've updated the bug with an example of how this breaks fts tables
> (fts1 or fts2). I'm thinking on the problem.
> http://www.sqlite.org/cvstrac/tktview?tn=2510
>
> Summary: In sqlite 3.4
You can, and I'm working on a patch to do this to see how it might look.
There's the question of how to handle existing tables.
-scott
On 7/17/07, Chris Wedgwood <[EMAIL PROTECTED]> wrote:
On Tue, Jul 17, 2007 at 09:37:43AM -0700, Scott Hess wrote:
> Summary: In sqlite
7;ll try to add more constraints to the summary today,
-scott
On 7/17/07, Scott Hess <[EMAIL PROTECTED]> wrote:
WTH! Wow, this is a very unexpected change. I must have not been
paying attention at some point.
-scott
On 7/17/07, Ralf Junker <[EMAIL PROTECTED]> wrote:
>
&g
WTH! Wow, this is a very unexpected change. I must have not been
paying attention at some point.
-scott
On 7/17/07, Ralf Junker <[EMAIL PROTECTED]> wrote:
>>The standard way to have non-TEXT information associated with rows in
>>an fts table would be a separate table which joins with the ft
On 7/16/07, Adam Megacz <[EMAIL PROTECTED]> wrote:
Is there any way to use a INTEGER PRIMARY KEY AUTOINCREMENT on a table
that has FTS2? Specifying it in the obvious manner looks like it
works, but the column just ends up with nulls in it.
In fts tables all columns other than rowid are of type
On 7/11/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
"Scott Hess" <[EMAIL PROTECTED]> wrote:
> do_test test-1.1 {
> execsql {PRAGMA encoding}
>
> sqlite3 db2 test.db
> execsql {CREATE TABLE t (id int)} db2
> db2 close
>
> #execsql {SEL
Aha, yes, that does sound like exactly it. Thanks!
On 7/11/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
--- Scott Hess <[EMAIL PROTECTED]> wrote:
> do_test test-1.1 {
> execsql {PRAGMA encoding}
>
> sqlite3 db2 test.db
> execsql {CREATE TABLE t (id int)} db2
&
do_test test-1.1 {
execsql {PRAGMA encoding}
sqlite3 db2 test.db
execsql {CREATE TABLE t (id int)} db2
db2 close
#execsql {SELECT * FROM sqlite_master}
catchsql {SELECT * FROM t}
} {1 {no such table: t}}
Looks like the schema info is obviously being cached. The first
PRAGMA is to make su
On 7/11/07, Steve Krulewitz <[EMAIL PROTECTED]> wrote:
In the application I am working on (Songbird), we have a simple two
table schema representing the tracks in your music collection and the
properties on those tracks. The keys used are all UUIDs (128 bit
number) which we are currently storing
On 7/10/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
--- [EMAIL PROTECTED] wrote:
> "Scott Hess" <[EMAIL PROTECTED]> wrote:
> > If you have an fts1 table f, you could drop f_term and f_content, but
> > you won't be able to drop f itself. So you wo
27;re Hiring! Seeking a passionate developer to join our team building
products. Position is in the Washington D.C. metro area. If interested
contact [EMAIL PROTECTED]
On 7/9/07, Scott Hess <[EMAIL PROTECTED]> wrote:
>
> If you have not compiled in fts1, and try to drop an fts1 table,
If you have not compiled in fts1, and try to drop an fts1 table,
you'll get an error. I don't think you'll get a crash, but sqlite
will simply not know how to deal with the table.
I can think of two ways to deal with this. When builidng a new
version, you could just leave the fts1 code in place
In the misty mists of time, we made the decision to minimize namespace
pollution by trying to keep everything in a single file. fts2.c is
almost to 6000 lines, almost as big as btree.c. Does anyone have
strong thoughts about the negatives of breaking things up?
Offhand, the major bits of distin
On 6/10/07, Mark Gilbert <[EMAIL PROTECTED]> wrote:
I have just started learning about Full Text search in SQlite, and I
have some questions
Sorry for the delayed response, I just noticed this email waiting for
an answer, and nobody has ...
1) With the Amalgamated Sqlite, I guess to enable F
Recently, various incremental vacuum features have been added, so you can do:
-- Turn on incremental vacuum.
PRAGMA auto_vacuum = incremental;
-- Release 100 pages from the freelist, or all if there are fewer.
PRAGMA incremental_vacuum(100);
-- Return the number of items on the freelist.
On 6/20/07, Andrew Finkenstadt <[EMAIL PROTECTED]> wrote:
How difficult do you think it would be to support an alternative method of
indexing within SQLite specifically to support O(1) retrieval of the rowid
for a table, and then potentially O(1) retrieval of the row data for a
table, when in-ord
On 6/18/07, Sean Cunningham <[EMAIL PROTECTED]> wrote:
There was talk in the mailing list a while back about creating a new
operator that would act as a superset of '==' which would treat
NULL==NULL as True. I have seen this in some other database.
Anybody know if this is on the roadmap?
It wo
You can use something like:
select tableA.path, tableA.value from tableA,tableB where
tableA.path=tableB.path and (tableA.value=tableB.value or
(tableA.value IS NULL AND tableB.value IS NULL));
It's possible that won't use an index, either, due to the OR, in which
case you could try a union betw
On 6/14/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
You can't infer a function's return type from its arguments.
Take the hypothetical function FOO(x). If I pass it a number, it will
return the number spelled out as TEXT, but if I pass it a BLOB it will
return its length*PI as a FLOAT.
Would it b
On 6/10/07, Ralf Junker <[EMAIL PROTECTED]> wrote:
Hello Scott Hess,
>I've lined up some time to work on fts, again, which means fts3. One
>thing I'd like to include would be to order doclists by some baked-in
>ranking. The idea is to sort to most important items
I've lined up some time to work on fts, again, which means fts3. One
thing I'd like to include would be to order doclists by some baked-in
ranking. The idea is to sort to most important items to the front of
the list, and then you can do queries which limit the number of hits
and can thus be sig
On 6/7/07, John Stanton <[EMAIL PROTECTED]> wrote:
I have just started to use FTS2 and it is working well but I would like
to ask any other users if they have had good or bad experiences and why
they would use FTS2 rather than FTS1. The software is new and I have
not seen any feedback at this st
In case it wasn't obvious, the "more complicated way" would probably
be something like:
BEGIN;
ALTER TABLE RENAME MyTable TO MyTableOld;
CREATE TABLE MyTable (
SameColumn INTEGER,
NewColumn TEXT
);
INSERT INTO MyTable SELECT SameColumn, OldColumn FROM MyTableOld;
DROP TABLE MyTableOld;
COMMIT;
On 5/1/07, Samuel R. Neff <[EMAIL PROTECTED]> wrote:
One suggestion though, instead of (or in addition to) using '*' as the
prefix operator perhaps '%' would be more appropriate in order to be closer
to the LIKE operator.
Hmm. I was mainly just doing what other groups appear to do (Lucene,
MYS
I just finished ( http://www.sqlite.org/cvstrac/chngview?cn=3893 )
checking in a string of changes to fts2.c to provide prefix search.
This works like:
CREATE VIRTUAL TABLE t USING fts2(c);
INSERT INTO t (c) VALUES ('This is a test');
INSERT INTO t (c) VALUES ('That was a test');
INSERT INTO
Additionally, note that if you use ORDER BY, and it _is_ in the
indicated order already, then sqlite will optimize the ORDER BY away
entirely. So use ORDER BY.
-scott
On 4/17/07, Andrew Finkenstadt <[EMAIL PROTECTED]> wrote:
The order of the rows returned by a select that does not have an ORD
On 4/13/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
"Scott Hess" <[EMAIL PROTECTED]> wrote:
> My assumption would be that you'd use 'cvs log VERSION' to find the
> datestamp where the VERSION file was updated, then check the tree out
> using -D
My assumption would be that you'd use 'cvs log VERSION' to find the
datestamp where the VERSION file was updated, then check the tree out
using -D.
Right?
Thanks,
scott
-
To unsubscribe, send email to [EMAIL PROTECTED]
On 4/11/07, Dan Kennedy <[EMAIL PROTECTED]> wrote:
On Tue, 2007-04-10 at 09:26 -0700, Scott Hess wrote:
> On 4/10/07, Dan Kennedy <[EMAIL PROTECTED]> wrote:
> > I checked the code and conflict handling mechanisms (OR ERROR,
> > OR ABORT, OR REPLACE) do
On 4/10/07, Dan Kennedy <[EMAIL PROTECTED]> wrote:
I checked the code and conflict handling mechanisms (OR ERROR,
OR ABORT, OR REPLACE) do not apply to virtual tables.
Something to think about anyhow... Do we want conflict handling
for FTS (and other virtual modules)?
I think OR REPLACE woul
Thanks for the concise report. I'm going to take a look at this
today, to see if it's an fts1/2 problem. If it's _not_, I'll still
look at it, but perhaps with less eventual success :-).
-scott
On 4/9/07, Paul Quinn <[EMAIL PROTECTED]> wrote:
Very simple to replicate:
CREATE VIRTUAL TABLE
An interesting approach would be to use some sort of async I/O
facility to implement read-ahead.
Short of that, I have found that in some cases, on some operating
systems, implementing explicit read-ahead buffering for fts2 segment
merges improves performance when the disk caches are cold. Linux
I don't see how your modified version is any better than just putting
the sqlite3_bind_int() inside the loop. You've superficially lifted
some code out of the loop, but sqlite3_step() is going to have to go
through and bind all of the "pointer bound" variables in your
suggested API, so it won't s
Are you using explicit transactions at all? If not, as a quick test,
put the _entire_ job in a transaction and see what happens.
-scott
On 3/15/07, Ken <[EMAIL PROTECTED]> wrote:
To answer your question:
Yes I can use a flat file at this stage, but eventually it needs to be
imported into so
I took Makefile.linux-gcc, and made the obvious changes (there are
mingw lines all over in there). EXE = .exe, SO =dll, SHPREFIX = ,
[that was nothing for that setting], and TCC, AR, and RANLIB set to
the path to the appropriate commands from mingw.
I should warn that I haven't actually built us
On 3/13/07, Ralf Junker <[EMAIL PROTECTED]> wrote:
Scott Hess wrote:
>Keeping track of that information would probably double the
>size of the index.
With your estimate, the SQLite full text index (without document storage) would
still take up only 50% of the documents' si
On 3/13/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
Ion Silvestru <[EMAIL PROTECTED]> wrote:
> To Ralf:
> >As a side effect, the offsets() and snippet() functions stopped working,
> >as they seem to rely on the presence of the full document text in the
> >current implementation.
>
> Did you
On 3/10/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
There's also this - CreateFileTransacted():
http://msdn2.microsoft.com/en-us/library/aa363859.aspx
Doesn't this require Vista?
-scott
-
To unsubscribe, send email to
On 3/9/07, Mitchell Vincent <[EMAIL PROTECTED]> wrote:
I'm looking into ways of changing the schema type name when I do my
SQLite2->SQLite3 conversion. Most likely I'll have to pipe the .dump
from the SQLite2 db through a program to replace 'varchar' with
'text'... I'm working on that now!
You
e of .dump
would be .read.
-scott
On 3/9/07, Scott Hess <[EMAIL PROTECTED]> wrote:
On 3/9/07, Gunnar Roth <[EMAIL PROTECTED]> wrote:
> Anderson, James H (IT) schrieb:
> > I need to "export" a table to a file in the same format as used by
> > .import, but I do
On 3/9/07, Gunnar Roth <[EMAIL PROTECTED]> wrote:
Anderson, James H (IT) schrieb:
> I need to "export" a table to a file in the same format as used by
> .import, but I don't see any such cmd. Am I missing something, or does
> such a cmd just not exist?
Maybe its dumb but its called .dump ;-)
S
I think only the dot commands are special (.help, etc). Everything
else is fair game. Best reference for what you can feed a prepare or
exec is http://www.sqlite.org/lang.html .
-scott
On 2/23/07, Igor Tandetnik <[EMAIL PROTECTED]> wrote:
Stef Mientki <[EMAIL PROTECTED]> wrote:
> Igor Tandet
Fix is in
http://www.sqlite.org/cvstrac/tktview?tn=2219
[This covers both fts1 and fts2.]
-scott
On 2/6/07, Scott Hess <[EMAIL PROTECTED]> wrote:
http://www.sqlite.org/cvstrac/tktview?tn=2219
ATTACH DATABASE 'test2.db' AS two;
CREATE VIRTUAL TABLE two.t2 USING fts2
http://www.sqlite.org/cvstrac/tktview?tn=2219
ATTACH DATABASE 'test2.db' AS two;
CREATE VIRTUAL TABLE two.t2 USING fts2(content);
will put t2_content, t2_segments, and t2_segdir in database 'main'
rather than database 'two'. In many cases everything will appear to
work, because the tables wi
301 - 400 of 428 matches
Mail list logo
