ter -
instead you need to measure and tweak multiple places.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlOGMfgACgkQmOOfHg372QT9IACfVvhc1LWG4X2IFBC0rKKNnrdw
UFIAoNhtFdh1EZKEo3fx7Kj9bkdKJRW4
=02fs
-END PGP SIGNATURE-
___
size), or by internal compression with cerod:
http://www.hwaci.com/sw/sqlite/cerod.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlOFpvsACgkQmOOfHg372QRSRACfcDqTprcD//n9yYXcGPl9yQfo
sTIAoLkIaQHR4JAwk1LbuRzCyQsx/5aN
=tYeT
-END P
e years:
http://www.sqlite.org/src/tktview?name=25e09aa2ab
The SQLite team rejected it in March.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlN/2XsACgkQmOOfHg372QRtnACeOCNixsgztnpZSqk/4T4fWqZE
fN4An0t4mC2VQOXwXeCnV3qYrUTULHRx
=qk0x
-E
utside bit was
because of having to do UTF8 conversion from native string representation,
and inside because statement preparation takes a while - it involves
parsing, many memory allocations and lots of other fiddly stuff.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlN7rRM
have to have a copy of the SQL statement.
SQLite already has to deal with all these issues, including keeping a
copy of the statement so SQLite implementing the cache would save memory.
Roger
___
sqlite-users mailing list
sqlite-users@sqlite.org
http:/
There is a ticket including pointers to previous mailing list discussion:
http://www.sqlite.org/src/tktview?name=ee4b2b48f5
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlN6dfEACgkQmOOfHg372QQmcACfX46ZvXwKt/Y9DnBL0BRxH9QM
lFsAoLRh/hjZai4SC5ie/DCpd2+D3NA2
=1Ve9
m the public API.
SQLite could provide the information as a virtual table or similar. A
ticket was created 7 years ago asking for it, and closed 2 months ago by
the team with resolution "Rejected":
https://www.sqlite.org/src/tktview?name=5896edbe46
Roger
-BEGIN PGP SIGNAT
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 22/04/14 15:00, Neville Dastur wrote:
> On 22 Apr 2014, at 21:58, Roger Binns wrote:
>> Your data is from MongoDB :) Note they do have an extended JSON to
>> deal with types like ObjectId, binary and dates:
> Yes, it is.
tored. You'll have to
write them to understand your data shape and queries.
There was a now defunct project UnQL that was mixing together JSON like
data, SQL like queries and bit of SQLite.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlNW19UACgkQmOOfHg372QTNvgCgltU2vNzzVuRfTuPd
out dependency order.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
iEYEARECAAYFAlNP58cACgkQmOOfHg372QT0RwCg5gXqYYmQ0Ws1LqJaFBhdpvwJ
dLAAn2+8hvf+nl9jTR45Z2TYQsT6mAlB
=APvL
-END PGP SIGNATURE-
___
sqlite-users mailing list
sqlite-
erything up so Python being used was not visible (nor relevant) to the
user. This approach worked fine on Windows, Linux and Mac.
[1] distutils - a standard part of python - is used under the hood which
has multiple subcommands each of which can be given flags. pip etc
ultimately call into that.
nail down
the cause for.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU/Linux)
iEYEARECAAYFAlMvwWMACgkQmOOfHg372QSnOQCdEpBWBvcNsntkZ6WPvDs0yAju
fc0AoJzagj56DyoYrhmeE73rwHhe+D2f
=ZAfw
-END PGP SIGNATURE-
___
sqlite-users m
the data, just the structure of the data. There was
a feature request ticket for several years for checksums to at least catch
unexpected changes to the data itself:
https://www.sqlite.org/src/tktview?name=72b01a982a
Sadly it was rejected a few weeks ago without explanation.
Roger
-BEG
levant code including any
code that calls that code, to audit for overflows, to provide a new api
and tests for that.
And to ensure that stays maintained for the future lifetime of SQLite 3.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU
aspx
You can mitigate it by having a setup that doesn't have failures such as
using battery backup.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU/Linux)
iEYEARECAAYFAlMUxBMACgkQmOOfHg372QTxewCgjuqKWh4m+pz2JRtQWznPA83o
YEcAnjDuMMULpMX14VVlLsQ4NmJbD6PA
=Dp0Y
.@126.com
http://catb.org/~esr/faqs/smart-questions.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU/Linux)
iEUEARECAAYFAlMSPxoACgkQmOOfHg372QQXIQCfT7Sa9kempGXWTYs+L6aot98I
Rw8AmJ2dp9jR0bN0FThp98ab/ZygeD0=
=T0I3
-END PGP
d the files I would
> greatly appreciate it I think the file ext. is a plist. Live, love &
> laugh.
In addition to those, read this:
http://catb.org/~esr/faqs/smart-questions.html
Roger
-BEGIN PGP SIGNATURE-
Vers
actually have. It also supports very old versions of
SQLite - I believe it will work correctly with 3.0.0!
> If we wish to have SQLite + Python combination
You'll find the python sqlite community at
https://groups.google.com/forum/#!forum/python-sqlite
Roger
-BEGIN PGP SIGNATURE
with other
files/directories/data in flight.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU/Linux)
iEYEARECAAYFAlLYT00ACgkQmOOfHg372QTw3QCfQNoOjJHrs9+pTSNmAx5BCoXk
essAoM0wmW454CL54bf0u2XTqV9vNxXP
=DY1e
-END PGP SIGNATURE-
___
sqlite-us
nularity of one second. The blocking thread could have
finished after 10ms, but you'll still be stuck in the busy handlers for
another 990ms.
Simply ensure HAVE_USLEEP is defined when building sqlite3.c. Or add your
own busy handler that sleeps for sub-second amounts of time.
Roger
n that SQLite should have a statement cache built in
rather than everyone having to implement their own:
https://www.sqlite.org/src/tktview?name=ee4b2b48f5
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU/Linux)
iEYEARECAAYFAlLB3aYACgkQmOOfHg372QSsFgCgwLkmzGqOrVL9YRLy6gG1
column: nonsense
sqlite> select nonsense and 0;
0
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.14 (GNU/Linux)
iEYEARECAAYFAlKzo/8ACgkQmOOfHg372QTyVACfbNlMOQR5nxfzgUg7tY5pF8il
d0sAnjIp9K/aAIFXZcbc+pZvDdK6tN0x
=Siia
-END PGP SIGNATURE-
__
detect
changes since you last looked at the file.
You can use a trigger internally at the SQL level to track changes in a
meaningful way too.
Finally you can disable the use of memory mapping - have a look at the
Windows VFS.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4
for many years to at least make
detecting a mismatch between the database and a (possibly non-existent)
journal possible:
https://www.sqlite.org/src/tktview?name=61d35ac210
Roger
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:808
brainz
http://musicbrainz.org/doc/MusicBrainz_Database/Schema
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlJ8ZWQACgkQmOOfHg372QQN/wCg0EyCpIs8pl9MVOzerfshJKxc
JwAAoN+E6Wa8T3zWZ+r5ghLpeK1a8dem
=23xE
-END
nd logCat shows the SQLITE_BUSY
> error code.
On Android I use the SQLiteDatabase class, and the insert method it
provides and haven't encountered this issue.
https://developer.android.com/reference/android/database/sqlite/SQLiteDatabase.html
How are you inserting the data?
Roger
-BEGIN
id.com/reference/android/content/Context.html#getCacheDir()
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlJTb3YACgkQmOOfHg372QTbUwCguSVoQ7AcJrGCk7qwwhzFQOqQ
z5YAoIFNQxhOTeP8/0Aq2lEpOD0Lc+GT
=VUgP
-END PGP SIGNATURE-
API.
http://apidoc.apsw.googlecode.com/hg/pysqlite.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlJCZOcACgkQmOOfHg372QRaYwCgqoR4C9DmXg27SCbOXg2a4GNe
vyUAnRYFf8Uej21hrJTkhwXJWddkvYWi
=bKyX
-END PGP SIGNATURE-
__
(eg CoreData on Mac likes to do that). As another bonus
it is also a bit faster too as the compiler ends up inlining SQLite code
into your methods that call it.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlJCBoIACgkQmOOfHg372QRRsQCePzIDcnfY
you gave in your example, truncating to 30
characters, and whatever else is relevant for your data. You can now do
matching against the normalised title column for each title.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iE
SQLite go off and do whatever
it deems necessary to gather to make the query perform well in the future.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIyQagACgkQmOOfHg372QT8jgCgtSROjcL1dyrHo+yP2leh1ffV
xBEAoKEOTIVqz3vlrVrlVeJ130Wru/Mg
=+8TU
-
that pysqlite parses your SQL and tries to second
guess what is going with transactions, and starts them behind your back.
That is possibly causing problems for you too.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIk8J8ACgkQmOOfHg372QTYCQCgi79D27GRMc6Tgjc
elease.
Note that the APSW API is similar to pysqlite, but not the same. It
behaves the SQLite way rather than the DBAPI way.
http://apidoc.apsw.googlecode.com/hg/pysqlite.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIj0vYACgkQmOOfH
t can be caused by other rare sequences of api calls.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIj0AIACgkQmOOfHg372QTU5QCfeGcle8r3m8nDs3dDgI7zb6BK
qWQAoLWkeKarpubX4S1SzcBxaT5LT1yr
=sjZY
-END PGP SIGNATURE-
_
bout this for several years.
http://www.sqlite.org/src/tktview?name=c060923a54
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIXpJkACgkQmOOfHg372QQDFQCgn56WYKUg3oxbiZ0Kp886flWa
H+sAoLMWH6Q462K1Lb5pTs7WwlZyxlnf
=Uj9e
-END PGP
, which is why it turns out to be the least exercised
part of protocols and where bugs/quirks lie.
[1] http://www.rogerbinns.com/visionfs.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIMB3wACgkQmOOfHg372QTyGgCgxB/7XxUibw+p0T/aI0cNecE4
ZegAoMDoh7RKb6h
tc can also generate arbitrary
messages.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlIIrLsACgkQmOOfHg372QS2pACeK5gt9bOdk9dX3gM+kxhYt6xA
MxQAnRLb4yMHdNEV6RpV0MbFUQVteNJ/
=gJrr
-END PGP SIGNATURE-
___
sqlite
results of EXPLAIN QUERY from
> sqliteman.
Attachments get stripped from the mailing list. You can put them
somewhere like Dropbox.
Did you run analyze?
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlICmtMACgkQmOOfHg372QSRLwCgkb
thonic than the
SQLite wrappers (disclosure: I'm the author of one of them). Code can use
iterators and look something like this:
for id,one,two in db.cursor().execute("select id, one, two from XXX"):
three=one+two
db.cursor().execute("insert into ", (three,
e you can immediately rerun the statement with the
same bindings. However that isn't exposed in APSW so you can edit the
code to clear the bindings after each reset call.
It doesn't do that at the moment as a form of being lazy - only clearing
bindings when the statement is re-executed.
R
te3_finalize is only called when the statement is being destroyed
which includes eviction from the statement cache, and on getting
SQLITE_SCHEMA error.
I do wish SQLite had a standard statement cache available. Pretty much
all wrappers implement their own which is duplicate work and tricky to get
right
ut would be ugly.
It will be interesting to see who wins the query syntax mindshare. There
does seem to be a desire to stay with SQL like syntax - for example
Cassandra also uses it. Another thing I like about the Mongo syntax is
that it is trivial to build up queries programmatically.
Roger
-BEGIN PG
data are
substantially similar - it is very similar to Query By Example.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlHaGJwACgkQmOOfHg372QTXaQCgkiy49SE5bHyvRr8gn+Oecgve
kSMAoLcq1ZXRg4Z5KXdU3HsfT9CkyqhE
=J/3c
-END PGP SIGNATURE-
_
???', '') as x from a;
1
sqlite> select instr('', '') as x from a;
3
sqlite> select instr('????', '') as x from a;
5
sqlite> select instr('', '') as x from a
599k mark.
You probably want -Os
http://www.sqlite.org/footprint.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlHUyqQACgkQmOOfHg372QTkQACgkKeNo1Qk5G4sgdkLJZPKV/Ai
0nkAn34ULSWYCsNk6I5WHvDViQOeTRTJ
QLite. It is already designed and implemented to
do the least amount of work - you haven't found a way to do less because
SQLite is already doing the least!
It only calculates the next (or first) result row when you ask for it, and
not before.
Roger
-BEGIN PGP SIGNATURE-
Version
xtensions:
http://msdn.microsoft.com/en-us/library/windows/desktop/aa378870(v=vs.85).aspx
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlHN+ugACgkQmOOfHg372QTTqgCeN3hNpHGON/CaoEx95y9605Qz
GMIAnAtmuJzgo3wwLrZdGOIbA2yWwuRP
=5SeP
-
create table ""("" "");
If your code runs on a platform that is UTF-16 (or worse UCS-16) make sure
you test with code points above 65535.
For performance testing this could be helpful with the --dump-sql arg:
https://code.google.com/p/apsw/source/browse/tools/
predictable so the query solutions chosen tend that way.
An upcoming release will have a new query planner that should work better
when there are a lot of joins.
http://www.sqlite.org/draft/queryplanner-ng.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4
the latter.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlHB29kACgkQmOOfHg372QTxwQCfc3qGW5PtGljBm/rvgEOGVtA/
DzwAoI6cIiZ+SzMmX7NE2gcwEfDR0w2W
=y1z6
-END PGP SIGNATURE-
___
sqlite-users mailing list
sqlite-users
e has not had any changes in two years.
Also if you want to use SQLite and not pretend that it is the same as
other databases then this will be of interest:
http://apidoc.apsw.googlecode.com/hg/pysqlite.html
(Disclosure: I am the author of APSW)
Roger
-BEGIN PGP SIGNATURE-
Version: GnuP
;t used:
http://apidoc.apsw.googlecode.com/hg/execution.html#apsw-trace
There is a sqlite3_trace function but I don't use it because it expands
bindings which means you can't lookup the queries in your code.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEAR
Would this not work for you?
http://stackoverflow.com/questions/7426205/sqlite-adding-comments-to-tables-and-columns
.schema
On 13-06-14 11:00 AM, Dave Wellman wrote:
Thanks Clemens, that is probably a workable option (at least for me).
As someone else noted, the PRAGMA user_version will no
SQLite then decides to use the index you can then
create it on the fly.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlGuQQMACgkQmOOfHg372QRxxwCfV1QvfO6Fsky0x3krTLe08+Nm
EN4AniM6kkOHgcTm/mtREY3iD4QAF9o+
=
ection phase.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlGqUEUACgkQmOOfHg372QRVXACdG1ZDSUzrL2TGMfaE1cQuSUgN
6vMAmQHwoWKQ+/zhAuZUNwM2U2Om4SAB
=P6JF
-END PGP SIGNATURE-
___
sqlite-users mailing list
sqlite
r each column while the second pass actually
imports the data.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlGpIoEACgkQmOOfHg372QRSqwCg3WTRwifRKC+tK55BmTbomVyo
PHsAnR8v79nKkpVZ7WYWydeTYxsHbZoE
=c95H
-END PGP SIGNATURE-
_
even show up under pragma check.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlGDT0QACgkQmOOfHg372QSkKwCgysgQWSsCWgSN1zMS0p9uOOWb
TL8An1MuVaYjysden5anmCnKk2G3+b5w
=8HDa
-END PGP SIGNATURE-
___
sqlite-users m
an
change bits in memory. Filesystems without checksums can have bit flips
on the storage or in transmission.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.12 (GNU/Linux)
iEYEARECAAYFAlGDTf0ACgkQmOOfHg372QSuhgCfZaiQPow1ioBbZnez569/oVPt
3s0AoLdwcTPuEB8apk5Dv1VNpLS2bMXC
=VYvk
---
w.c#217
Here is the corresponding test:
https://code.google.com/p/apsw/source/browse/tests.py#7562
(I use 1/0 to ensure that there will always be an exception, and if it
turns out to be ZeroDivision then I know the previous line didn't cause
one and should have.)
Roger
-BEGIN PGP S
This is discussed in:
http://www.sqlite.org/c3ref/blob_open.html
It mentions using zeroblob() so you can create the blob of the length you
want and then use blob_open/read/write/close to modify the contents.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (
http://www.sqlite.org/howtocorrupt.html (section 2.1)
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlFp78AACgkQmOOfHg372QRcWQCfcPTuzBjSGNUeYR6xwQSFyvDe
G8kAnim8px5ZTDow9KcbqebHBu3Nod/2
=llAx
-END PGP SIGNATURE-
_
The
reason is because I can put arbitrary JSON objects in and get the same
arbitrary JSON objects out. Queries have the same "shape" as the JSON
objects (unlike the defunct unql which tries to use SQL).
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.
ir mangled data and ids to make future matching easier
- - almost every system like this ends up "intelligently" duplicating all
the user records at some point. Think about prevention and how the user
recovers (undo from earlier is a great help)
Roger
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
http://www.sqlite.org/lang_corefunc.html
The hex line is out of alphabetical order and should be two lines earlier.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux
ables, with triggers for insert to redirect columns as appropriate
- - convince/pay someone to make a patch that implements what you need and
maintain your fork forwards
- - convince/pay the SQLite team to add it to the core, keeping in mind that
they care about library size and the '
ows better.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlFGzgAACgkQmOOfHg372QSzoQCfWxH+ACpo3FbeYFli7DEZVGpO
G+IAn2twe1Lva3sVE05jedQi/7VL3lpk
=CFQH
-END PGP SIGNATURE-
___
sqlite-users mailing list
sqlite-us
te data. My test suite extensively covers the
SQLite C API. Everything looks good and I don't even remember the last
release there was a problem.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlE/2oUACgkQmOOfHg372QS3CwCfY/P+4IM
SQLite thinks it wrote is not what is
later returned:
http://www.sqlite.org/src/tktview?name=72b01a982a
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlExZhYACgkQmOOfHg372QTLBgCgsbidy6oQfmAeS4OWq4OBSmFI
zxEAn04lneghgvr+ww76AQWzycZ3x+Q
ing TEXT, sync_status INTEGER, _natsort_name TEXT
COLLATE NOCASE);
CREATE TABLE pending_uploads (_id INTEGER PRIMARY KEY AUTOINCREMENT, class
TEXT, data TEXT);
dropbox._data is the path to a local copy of the file, not the file contents.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (G
TLDR: good question and there are a *lot* of issues under the surface
depending on how you intend to do the restore and what things matter to
your code
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEwsFoACgkQmOOfHg372QSsTQCffwvaQFk1c/XAhZHJpPxCHflE
2PUAn1R
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 24/02/13 00:14, Roger Binns wrote:
> On 23/02/13 21:23, Ashok Pitambar wrote:
>> Sqlite doc says it thread safe ...
>
> http://www.sqlite.org/threadsafe.html
>
> What makes you think the doc is wrong?
One big gotcha: er
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 23/02/13 21:23, Ashok Pitambar wrote:
> Sqlite doc says it thread safe ...
http://www.sqlite.org/threadsafe.html
What makes you think the doc is wrong?
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAl
up in the operating system
cache and so perform at the speed of memory (no seeks, no disk).
http://en.wikipedia.org/wiki/Amdahl's_law
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEmTxAACgkQmOOfHg372QSq/QCgu6iT0Y+XBpLyczh
append "';". (That is essentially what
the .import command is doing behind the scenes, but it tries a lot harder
to find separators, count/enforce number of columns, deal with end of line
etc.)
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEAR
uot;.
http://www.sqlite.org/src/info/25e09aa2ab
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEYS6EACgkQmOOfHg372QR+kgCePgJMHJ55coumXdEIsIOlOhJr
Ns8An1bmLOFeJ/5vYMBa8nQJSw31MmLC
=czW1
-END PGP SIGNATURE-
___
sqlite-
ck=full --leak-resolution=high --show-reachable=yes
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEWueoACgkQmOOfHg372QQtaQCeNfCzwMp2UzNslMjoI538tjBf
0SEAoNEfzvG70jqtWeY7T2LcVfjcFYAM
=jmof
-END PGP SIGNATURE-
___
fault
(disclosure I am the author). The recommended build instructions embed
the SQLite amalgamation inside the extension so it doesn't affect anything
else.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlET+T8ACgkQmOOfHg372QQZw
tings but does work on every other
modern operating system.)
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEQxrAACgkQmOOfHg372QTlLgCeLI2dZP/v6Pw4VlgcgRGUKqHB
xcwAnjM1RAHNzGgvi9nnGnP4uoYG6vqV
=DCOk
-END
exist then you a row per column in
the index.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlELmyUACgkQmOOfHg372QSkCwCgvIzw59DcPMytdRssRs37YPRB
IUMAoKmyDMkKhdubaMqdmp/H9WLrkYAW
=WCNM
-END PGP SIGNATURE-
___
to the GIGO
> principle is perfectly adequate.
Note the change the SQLite team have made:
http://www.sqlite.org/src/info/b4d94947fc
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEHZEMACgkQmOOfHg372QTUTgCeOLdm5gpZQDdE6WIWZHZ
3(content /* ab */ TEXT)
Of course this is an implementation detail and not specified anywhere, but
it is good enough. Your like query has the problem that it will match
rows where virtual appears anywhere.
Roger
-BEGIN PGP SIGNATURE-
Version: G
. It looks like SQLite is the victim of some other memory
mismanagement in your app and changing the SQL has just changed what code
will fall victim to it.
It is a very good idea to run valgrind before proceeding.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1
#x27; 'hi ' 'hi t' 'i ' 'i t' 'i th' ' t' ' th' ' the' 'th'
'the' 'ther' 'he' 'her' 'here' 'er' 'ere' 're'
You can possibly al
OR desc LIKE "%red%" OR title LIKE
> "%soup%"...'
>
> Will creating a composite index ...
If you are concerned about performance then SQLite already has a solution
for you:
http://www.sqlite.org/fts3.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG
s
often in powers of two. It could be that row had a string needing 33
bytes while others needed less which then caused memory to come out of a
different bucket which then changes where the victim of the actual bug
shows up.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linu
C/ARC and C.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlEBu6sACgkQmOOfHg372QSR5gCfU+FAhhp8QuzAB6q1sP8xq2T4
j3wAoMUx3i4dFzzIldgyGJghNtXmsJR6
=+wFc
-END PGP SIGNATURE-
___
sqlite-users mailing list
sql
usleep).
{ 1, 2, 5, 10, 15, 20, 25, 25, 25, 50, 50, 100 }
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlD4erkACgkQmOOfHg372QQWPwCfZlMikT1DnuiflZDa58pIdnD9
vwYAoKRI+Z5UQWl6NP6rm8K1eyrRxYht
=vuw2
-END PGP SIG
:31
8609a15dfad23a7c5311b52617d5c4818c0b8d1e.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlDuIA0ACgkQmOOfHg372QS9yACfZFFwvjkqUhcDB0zXRiRzv8NK
LTsAn2nTH/sVr8BAHHogRixfXUfeQkVQ
=+1Ku
-END PGP SIGNATURE-
___
sqlite-users mailing list
sqlite-us
om/p/apsw/source/browse/src/apsw.c#704
In my benchmark tests I measured a slowdown of 1%. ie if your code did
nothing but SQLite calls then you can expect it to be about 1% slower.
I strongly recommend you do something like this to ensure that no
developer accidentally has databases used across f
e which
makes things slower.
Journaling off definitely worked for me when I benchmarked it. I was
working with a 15GB dump from postgres on Linux.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlDiABEACgkQmOOfHg372QQesQCgjTJbDb3Yt2iyC/7vUEJAFuTq
1T0AoIPDu/fpd
creating the initial database then you can turn journalling etc
off until the database, indices etc are fully created. This will get you
a little more speed too.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlDh7mkACgkQmOOfHg372QR/dwCfVhcMaYJIr6pTFKsL1LbaFiVJ
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 12/12/12 12:36, Richard Hipp wrote:
> Does anybody know of any reason why we should not do this?
I would love if this was combined with an optional per page checksum that
detects corruption early.
Roger
-BEGIN PGP SIGNATURE-
Vers
;ll probably find that quite hard. A simpler alternative is to
temporarily attach another database and copy the table into it.
attach 'backup.db' as backup;
drop table if exists backup.important;
create table backup.important as select * from important;
detach backup;
Roger
-
s/Riak-Search---Querying/
MongoDB uses JSON with extra magic operators:
http://docs.mongodb.org/manual/applications/read/
Personally I like MongoDB approach where queries strongly resemble the
underlying stored data which means very little translation between the two.
Roger
-BEGIN PGP SIG
ble is to use a virtual
table and do the backup during writes/sync.
This is all considerably more work than figuring out why you are getting
corruption in the first place.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEARECAAYFAlDGF2IACgkQmOOfHg372QTo9gCfSQQwreSvsa9lrV/
rom implementing undo by using triggers to save old
values, through using a virtual table that backs up data somewhere else.
The SQLite backup API only works on the whole database.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Li
C based process. Also if glib/evolution/libcamel play
silly games with memory (eg using their own allocators and pools) then it
will completely obscure what is going on from valgrind.
You'll most likely get better insight in a group devoted to evolution.
Roger
-BEGIN PGP SIGNATURE
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 02/12/12 13:47, Paul Menzel wrote:
> please allow signed messages to be sent to the list.
>
> I got: »The message's content type was not explicitly allowed«
This works.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
The documentation also needs to say asyncvfs is out of favour, in
particular on this page:
http://www.sqlite.org/asyncvfs.html
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.11 (GNU/Linux
27;t shouldn't using a current Windows
7/8 on a fresh machine today.
Run msinfo32 and then Components > Storage > Disks to find the relevant
partition and its starting offset/alignment.
(This is unlikely to be your problem, but if present does result in the
kind of performance degradation y
201 - 300 of 1622 matches
Mail list logo