Your SQL pass, so it is somewhere in the library. The problem is how to
find in the big data where exactly. I am now going one character by each
other, because it is probably only one specific record kind doing that.
I was able to compile the library under linux and running strace looks
like
So I was able to identify the record doing the problem. I have tried to
find what is so special on that record and it is size 20 characters.
When I have this 3 examples with last name lenght 19, 20 and 21
characters only the size 20 crash.
.load ./libstringmetrics.so
select a.firstname,
Hi Milan,
thx for your patch.
SQLITE_STATIC --> SQLITE_TRANSIENT
I try to apply it for testing:
but seem nothing change .
However I leave it in the code to allow you to test.
I submit also a new dll.
Regards,
Andrea.
2015-03-11 16:07 GMT+01:00 Milan Roubal :
> Dear Andrea,
> thank you for the
I have a table, Transactions, with a column , Value, of type NUMERIC. The
Value column is supposed to have 2 decimal places in all rows but some have
only one.
To correct this I issued the following command:
UPDATE Transactions SET Value=Value||0 WHERE substr(Value,-2,1)='.'
No errors on
> On 11 Mar 2015, at 10:54pm, Milan Roubal wrote:
>
> .load ./libstringmetrics.so
> select a.firstname, b.firstname, a.lastname, b.lastname,
> stringmetrics("qgrams_distance","similarity",a.firstname, b.firstname,"")
> first_dist,
> stringmetrics("qgrams_distance","similarity",a.lastname,
Hi Peter,
From https://www.sqlite.org/datatype3.html:
"When text data is inserted into a NUMERIC column, the storage class of
the text is converted to INTEGER or REAL (in order of preference) if
such conversion is lossless and reversible"; basically any fraction you
insert into a NUMERIC
On 2015-03-12 01:27 AM, Peter Haworth wrote:
> I have a table, Transactions, with a column , Value, of type NUMERIC. The
> Value column is supposed to have 2 decimal places in all rows but some have
> only one.
SQLite has no formatting inherent to the column value, there is no such
thing as
On 2015-03-12 00:33, Simon Slavin wrote:
>> On 11 Mar 2015, at 10:54pm, Milan Roubal wrote:
>>
>> .load ./libstringmetrics.so
>> select a.firstname, b.firstname, a.lastname, b.lastname,
>> stringmetrics("qgrams_distance","similarity",a.firstname,
>> b.firstname,"") first_dist,
>>
> On 11 Mar 2015, at 11:50pm, Milan Roubal wrote:
>
>> SELECT stringmetrics("qgrams_distance","similarity","Roubal",
>> "RoubalRoubalRoubalRo","");
> crash
Then that's the one to use for debugging. It is extremely simple and contains
no database access at all so now you know the problem has
Thank you very much for all feedback! the last example crashed also so I
have tried with try and error to trace it down into the library and it
looks like the problem are this 2 lines in file
src/libsimmetrics/simmetrics/tokenizer.c
tmp = calloc((init_len + qtype->qgram_len),
Hello everybody,
I have a situation where two processes are accessing the same SQLite
database. One process only reads from the database, one process reads and
writes.
These processes keep a single database connection open for the lifetime of
the process.
It seems to me that once the reader
Of course.
Thx for report and patch.
However I will contact the original author of the simmetrics library
https://github.com/jokillsya/libsimmetrics
to report him this error and the patch.
A.
2015-03-12 2:12 GMT+01:00 Milan Roubal :
> Thank you very much for all feedback! the last example
Hi Lina,
I tested your patch and it resolve the crash.
I submit the patch in the master and generate a new dll for windows.
Thx for patching.
A.
2015-03-12 8:48 GMT+01:00 Andrea Peri :
> Of course.
> Thx for report and patch.
>
> However I will contact the original author of the simmetrics
oops,
sorry for wrong typo.
"Hi Milan," is the correct.
:)
Again,
A.
2015-03-12 9:08 GMT+01:00 Andrea Peri :
> Hi Lina,
>
> I tested your patch and it resolve the crash.
>
> I submit the patch in the master and generate a new dll for windows.
>
> Thx for patching.
>
> A.
>
>
> 2015-03-12 8:48
> If those databases are small then running "PRAGMA integrity_check" on them
> should be very fast.
My application does that. This is how users find out that their settings
database has become corrupted.
So far it is just reported and then the config database is restored from last
working
> Why are you using shared cache, especially with WAL? Are your devices memory
> constrained?
I was under the impression that shared cache has performance benefits when used
in a MT environment.
I have enabled the error callback and it logs everything except SQLITE_SCHEMA
and SQLITE_OK.
In the log files I have received so far, none of these were reported (I use
special headers to identify SQLite errors because log files are often several
hundred MB in size).
I?m waiting for a
On Mon, Mar 9, 2015 at 10:38 PM, Scott Robison
wrote:
> A co-worker who is working on a project is interested in finding out if
> there is an effective ORM for C++ / SQLite. I've not used one so I'm
> turning to the list to see if anyone has a recommendation.
>
> Note: He's okay using SQLite
On 2015-03-12 04:38 AM, Barry wrote:
> Hello everybody,
>
> I have a situation where two processes are accessing the same SQLite
> database. One process only reads from the database, one process reads and
> writes.
>
> These processes keep a single database connection open for the lifetime of
>
Hi Richard,
Like said, we've already tried:
CREATE TEMP TABLE x AS (...query without outer sort...); CREATE xind ON
x (price); SELECT * FROM x ORDER BY 2
but it also takes 30 seconds; for some really esoteric reason, CREATE
TABLE x AS ... takes all of those 30s, even though the execution plan
One thing to add:
I was sometimes successful to remote-repair a corrupted database by telling the
user how to use sqlite3.exe and calling REINDEX.
After learning that, I added this to my diagnosis routine so if
integrity_check() returns something that?s not ?ok?, my application runs a
On 3/12/15, Mario M. Westphal wrote:
> One thing to add:
>
>
>
> I was sometimes successful to remote-repair a corrupted database by telling
> the user how to use sqlite3.exe and calling REINDEX.
>
>
>
> After learning that, I added this to my diagnosis routine so if
> integrity_check() returns
I dump the output to the log file so when a user sends me a log after the
diagnosis reported a ?repaired? damage, I see one or more entries like:
?row 2481 missing from index idx_settings_sndidmnun?
I will see if I can collect more log files in the coming months.
On 12 Mar 2015, at 2:51pm, Mario M. Westphal wrote:
> I dump the output to the log file so when a user sends me a log after the
> diagnosis reported a ?repaired? damage, I see one or more entries like:
>
> ?row 2481 missing from index idx_settings_sndidmnun?
>
> I will see if I can collect
Hello,
I know, I am touching a hot iron, when claiming that sqlite might have a memory
leak. But I am trying to debug memory leaks related to sqlite now since 4 days
and I am running against walls.
Here is my problem:
- I am writing an iPhone app using iOS 8.2, Xcode 4.2, sqlite 3.8.8.3.
-
On 12 Mar 2015, at 4:21pm, Matthias Schmitt wrote:
> - The memory leak appears in the iPhone simulator more often, than on an
> original iPhone device, but it appears always.
>
> - The leak appears with totally different type of queries. It might be, that
> a leak appears in a specific query
On 3/12/15, Matthias Schmitt wrote:
> Hello,
>
> I know, I am touching a hot iron, when claiming that sqlite might have a
> memory leak. But I am trying to debug memory leaks related to sqlite now
> since 4 days and I am running against walls.
> Here is my problem:
>
> - I am writing an iPhone
Thanks to all for explaining my confusion. The printf solution seems like
the best way to handle this since I don't need to worry about how many
decimal places are in the number.
On 2015-03-12 06:21 PM, Matthias Schmitt wrote:
> Hello,
>
> I know, I am touching a hot iron, when claiming that sqlite might have a
> memory leak. But I am trying to debug memory leaks related to sqlite now
> since 4 days and I am running against walls.
Hi Matthias,
Firstly, there are no
You should be also aware of a more common pitfall: unclosed result sets.
Any lock is held until you read PAST the last row or you call stmt_finalize
(or the equivalent abstraction in your DBAL). Always close select
statements.
On Mar 12, 2015 11:40 AM, "R.Smith" wrote:
>
>
> On 2015-03-12 04:38
There is also a more obscure side-effect exhibited by sqlite only: if
you issue consecutive unfinalized select statements, it never releases
the write lock (although they should be atomic), but it does trip the
deadlock mechanism; any writer in the wait will receive SQLITE_BUSY at
the very
Hello,
may i ask again ...
Can somebody help to successfully compile spellfix.c for windows?
Errors see below!
Cheers
sonypsx
-Urspr?ngliche Nachricht-
Von: sqlite-users-bounces at mailinglists.sqlite.org
[mailto:sqlite-users-bounces at mailinglists.sqlite.org] Im Auftrag von sonypsx
The zip should contain 4 files
shell.c
sqlite3.c
sqlite3.h
sqlite3ext.h
sqlite-amalgamation-3080803.zip
-Urspr?ngliche Nachricht-
Von: sonypsx [mailto:sonypsx at gmx.net]
Gesendet: Donnerstag, 12. M?rz 2015 18:56
An: 'General Discussion of SQLite Database'
Betreff: Re: [sqlite]
Hi,
I have a normal table, and a fts table that are joined by rowid's. E.g.
when I add a row to the normal table, I add some values into the fts
table as well, and set the rowid based on the normal table rowid (as
recommended).
Now I have restructured the normal table (i.e. delete column) using
Rael Bauer wrote:
> Now I have restructured the normal table (i.e. delete column) using the
> algorithm mentioned here a couple of times (insert into normal select
> (columns) from normal_OLD, etc..)
>
> This "resets" the rowid's of the normal table.
If you rely on the rowid values, you should
On 12 March 2015 at 20:20, Rael Bauer wrote:
>
> Hi,
>
.
.
.
> Now I have restructured the normal table (i.e. delete column) using the
> algorithm mentioned here a couple of times (insert into normal select
> (columns) from normal_OLD, etc..)
>
> This "resets" the rowid's of the normal table.
On 3/12/15, Barry wrote:
> On 13 March 2015 at 01:21, Dinu Marina wrote:
>
>> You should be also aware of a more common pitfall: unclosed result sets.
>> Any lock is held until you read PAST the last row or you call
>> stmt_finalize
>> (or the equivalent abstraction in your DBAL). Always close
>> Actually query one appears slightly faster,
>> Searching the PK index is faster as that is always a COVERING index.
>
>I was under the impression that the opposite is true, but I wasn't sure
>about that.
The primary key is only a covering index if you are only accessing fields
comprising the
Make sure that you have all the necessary .h files in the same directory as the
spellfix.c source code (the #include's use " " rather than < >, which means
look in the current directory, not the include path).
Use the following command line in the directory containing source (.c) and
header
39 matches
Mail list logo