Hi,
The MVStore should work well with Android. The main differences to SQLite
are:
* SQLite overwrites / writes blocks of the size 4 KB. With flash or solid
state disks this is not that great, as internally (by the firmware or OS)
those writes are translated to much larger writes, for example 128
Hi,
No, I'm sorry, I think there is currently no way to influence this.
Regards,
Thomas
On Monday, May 4, 2015, Peter wrote:
> I am having some problems with indices and order by in H2 1.3.176. I
> managed to get it to work when I only query one table by:
> * making sure the index has DESC s
Hi,
Yes, I think this is a bug. I'm still working on a test case.
Regards,
Thomas
On Tuesday, April 28, 2015, wrote:
> Hi,
>
> in my opinion, the code below contains a bug in the last line. Shouldn't
> it be
>
> *return ValueResultSet.get(x); *
>
> instead of
>
> *return ValueResultSet.get(rs
Hi,
I understand. However, I'm afraid I will not have time to work on this
issues. I think those uses cases are not that common. Patches or ideas on
how to fix this are welcome!
Regards,
Thomas
On Sunday, May 3, 2015, Vitali wrote:
> I would add also my case here.
>
> Seems a query like
>
>
Hi,
I think "smallint" doesn't use variable size coding, so with small values
(1, 2, 3), using "int" actually saves space. If you have randomly
distributed values within the range of smallint, then smallint uses less
space.
Regards,
Thomas
On Tue, May 5, 2015 at 7:45 A
Hi,
That's strange. The last part of the file is missing.
Looking at the H2 source code, the file truncate code is very simple and
conservative (MVStore.shrinkFileIfPossible, which is calling
getFileLengthInUse). It looks unlikely that there is a bug in this area.
Could you describe what you did
Hi,
I would probably use INT. Small values use less space than large values, as
variable size coding is used. Performance is not affected much; INT should
be a bit faster but not much.
Regards,
Thomas
On Saturday, May 2, 2015, Stig Christensen
wrote:
> Hi and thanks for a great database. I'm u
I get "
> java.lang.UnsupportedClassVersionError:
> org/h2/Driver : Unsupported major.minor version 52.0" trying to load the H2
> driver. Is there a Java compiler mismatch? We're using Java 1.7 64-bit.
>
>
>> On Apr 22, 2015 10:44 PM, "Thomas Mueller" > &g
Hi,
The following should work for an empty, and single element array:
select ();
select (1,);
But I agree it looks like a hack.
Regards,
Thomas
On Mon, May 4, 2015 at 12:07 PM, wrote:
> Hi Noel,
>
> your suggestion leads to the same result as cast (1 as ARRAY), the type of
> the array e
Hi,
> But already for many years the spatial support was provided by a
combination of geodb + hatbox libraries and integration in GeoTools world
Yes. However, those don't use the built-in R tree. Do they use an external
R tree?
> All these was done on BLOB type where a geometry WKB is stored.
Hi,
The database is trying to read past the end of the file... Would it be
possible to send me the database file to investigate?
Regards,
Thomas
On Wednesday, April 29, 2015, Osvaldas Ziukas
wrote:
> This happend with newest H2 version i got and updated programs. h2-1.4.187.jar
>
> In 3 Days i
Hi,
I wonder why the conditions are in "having" and not in "where". Also, I
would probably use "module in (...)" instead of that many "or" conditions.
But that shouldn't change the result.
We would need to have a complete test case, including "create table" and
"insert" statements.
Regards,
Thom
Hi,
Thanks for the patch!
I wonder, why do you need to log the session user, and not just the session
id?
I think I understand the problem you want to solve, but the patch is larger
than necessary: logAndConvert could be overriden for example, which should
make the patch a lot smaller. The patch
Hi,
H2 simply drops the constraint as well in this case. I understand this is
not how other databases work, this should be changed. H2 supports "drop ...
restrict" and "cascade", but currently ignores it. Patches are welcome!
Regards,
Thomas
On Monday, April 27, 2015, cmarquesumu wrote:
> I am
Hi,
Many users ran into problems because they used something like
"jdbc:h2:test" and then either didn't find the database file, or created a
second database when running the application in a different directory.
That's why in version 1.4.x, now relative path only work when using ".", as
in "jdb:h2
Hi,
Yes. The MVStore is what is beta. For that to be marked stable, I think we
would need at least a few weeks where no issues are found with this
storage. I would say were are almost there. The next release will still be
beta, and then I think the MVStore can be marked as stable (unless
something
; indexes.
> I fixed them and I created a push request for the change on Friday.
> If you have time would you be able to take a look at it?
>
> Thanks
> John
>
>
>
> On 17 April 2015 at 12:11, Thomas Mueller
> wrote:
>
>> Hi,
>>
>> That's
Hi Sim,
> Can't you create your own topic?
I don't understand. As far as I see, it _is_ a separate topic (email
subject).
> I have set a custom serializer for Java object contained in columns of
type OTHER
Why can't you use the default serialization mechanism? What problem do you
want to solve?
Hi,
Do you have any indexes? Did you read
http://h2database.com/html/performance.html ?
Regards,
Thomas
On Sun, Apr 26, 2015 at 6:52 PM, sim wrote:
> Hi,
>
> Let's assume I have two tables
>
> table clients
> =
> cl_id bigint primary key,
> cl_name varchar
>
> table requests
> ===
Hi,
Using Travis is a good idea, I will have a look at it.
In the past, I configured a Windows laptop to do a build (lasting about 3
hours, including code coverage) per night. This stopped working with
Windows 7, because automated "waking up from suspend" no longer works.
Currently, the automate
to tests things (testGrantSchema, at
the end), I fixed that.
I added an entry in the change log.
I updated the SQL syntax in the help.csv file.
Regards,
Thomas
On Thursday, April 23, 2015, Thomas Mueller
wrote:
> Hi,
>
> Thanks a lot! The patch looks good. It's great to have good test cases!
Hi,
Thanks a lot! The patch looks good. It's great to have good test cases!
I did make the following changes:
* Source code formatting / spaces: "} else {" instead of "}else{" and ") {"
instead of "){". To automatically find those places, I use Checkstyle, with
the configuration at src/installer
Hi,
Could you post the error message and stack trace please?
Regards,
Thomas
On Thu, Apr 23, 2015 at 1:25 PM, Rami Ojares > wrote:
> Hi,
>
> I have reported a related issue but maybe I can give now more clues.
> I am using 1.4.187 with mvstore and multithreaded true.
> However this issue I am
Hi,
Thanks for the patch! It looks good, I will commit it.
Regards,
Thomas
On Saturday, April 18, 2015, Jared White wrote:
> Greetings,
>
> When a LinkedTable is defined to point to a table in Oracle with a
> function-based index, H2 throws a NullPointerException when attempting to
> model th
Hi,
Yes, I will update the documentation.
Regards,
Thomas
On Saturday, April 18, 2015, sim wrote:
> Hi Noel,
>
> You was right, multi_threaded=true works fine. I can't see a thread
> blocking any more,
> but then h2 documentation has to be edited.
>
> Thank you.
>
> On Saturday, April 18, 201
> On Tue, Apr 21, 2015 at 10:59 PM, Thomas Mueller <
> thomas.tom.muel...@gmail.com
> > wrote:
>
>> Hi,
>>
>> Thanks a lot! Yes, I think it is a bug. I have a fix for it now. This is
>> related to a bugfix in version 1.4.179, change log "Referential int
Hi,
We cannot set A_ID to correct value (as part of insert) because this is an
> ID column for the actual value column "A".
> For example , column A could be "Customer Name" column.
> The A_ID column for this column would be "Customer_ID".
>
I would probably try to combine those steps, so that yo
Hi,
That's strange. I see that a null password, salt, or hash results in a NPE,
and I will fix that. But an existing database shouldn't have such users.
Could you send me the database file?
Regards,
Thomas
On Tuesday, April 21, 2015, Wes Clark wrote:
> I am using the latest build 1.4.187. I a
Hi,
As documented, for H2 the syntax is:
ALTER TABLE ALTER COLUMN x RENAME TO y
I'm not sure if Oracle supports that as well. As for rename, only renaming
the columns is currently supported:
ALTER TABLE tableName RENAME TO newName
Patches are welcome!
Regards,
Thomas
On Monday, Apri
Script attached. If you run in the browser console,
>> you can browse the schema and see the two indexes.
>>
>> On Thursday, April 16, 2015 at 10:43:47 PM UTC-7, Thomas Mueller wrote:
>>>
>>> Hi,
>>>
>>> Yes, I think there was a change in this are
Hi,
I suspect we will have to do something similar to ImageIO and do our
> own management.
>
I would like to avoid that. Instead, I would like to avoid temp files at
all. With the MVStore, there is only one case where temp files are used
(that I know of): very large transactions (updates, deletes
Hi,
I don't know what's wrong, but maybe your class is in a different package?
Regards,
Thomas
On Tuesday, April 14, 2015, blu10 wrote:
> Thanks!...
>
> Trying to test now
> I created a Trigger class that implements the Trigger interface. I
> compiled this to a class file and into a jar a
hanks again
> John
>
>
>
> On 16 April 2015 at 12:07, Thomas Mueller > wrote:
>
>> Hi,
>>
>> You are right, this is not yet working. I will committed a fix for it
>> shortly. However I'm not sure performance is good with many null entries (I
>> didn
Hi,
> You can have indexed column lower_x which will contain lower(x) and do
queries like
> ... where lower_x = lower(?)
Yes. One addition: the column "lower_x" can be a computed column, so you
don't have to update it yourself:
create table test(x varchar, lower_x varchar as lower(x));
i
it does not make sense for non-snapshotable datastores.
>
> I assume your answer will be related to MVStore maturity, right?
>
> Sergi
>
> 2015-04-16 21:20 GMT+03:00 Thomas Mueller >:
>
>> Hi,
>>
>> Thanks a lot for your help! I have merged the pull reques
Hi,
Most of those settings are dangerous, so I would expect you get a corrupt
database (at least with the PageStore). But if it's just a cache, then that
might be OK. I have to say that for others reading this email.
Version 1.4.x (the MVStore) doesn't use that many temp files. The only
situation
Hi,
Yes, I think there was a change in this area because of a bug, but I don't
fully remember.
Could you please create a simple test case that shows the problem? I mean
something like:
drop all objects;
create table parent(id int primary key);
create table child(id int primary key, parent_id int
Hi,
The problem seems to be that you use a slow algorithm, kind of like bubble
sort instead of quicksort. Your algorithm seems to take quadratic time
(twice the number of rows results in four times slower performance).
Why don't you use just one update?
update table XYZ set a_id = userDefine
Hi,
I have merged the pull requests.
> Next few months me and probably my colleagues from Apache Ignite will be
contributing to H2 often
That's great! By the way I'm an Apache committer as well (Apache Jackrabbit
Oak).
> it would be nice if pull requests will be processed faster.
Unfortunately
Hi,
Thanks a lot for your help! I have merged the pull requests.
I agree the bit field should not be needed. It is currently needed mainly
for the PageStore storage mechanism; I think this can be removed once we
only need to support the MVStore.
Regards,
Thomas
On Wed, Apr 15, 2015 at 11:38 AM
java:174)
> at org.h2.mvstore.db.MVTable.removeRow(MVTable.java:611)
> at org.h2.command.dml.Delete.update(Delete.java:94)
> at org.h2.command.CommandContainer.update(CommandContainer.java:78)
> at org.h2.command.Command.executeUpdate(Command.java:254)
> at org.h2.jdbc.JdbcStatement.executeIn
Hi,
Do you have a reproducible test case that I could run? What is the database
URL? Is this an old database or a fresh one?
Regards,
Thomas
On Thu, Apr 16, 2015 at 8:27 AM, Sourav Mandal > wrote:
> Hi,
> We have been using H2 databases in our production, recently due to
> one activity so
Hi,
Could you tell us, what database URL and what options do you use? Which
version was the database created with?
Regards,
Thomas
On Thu, Apr 16, 2015 at 9:05 AM, Noel Grandin > wrote:
> There are various possibilities:
>
> (1) Don't use SET LARGE_TRANSACTIONS TRUE
>
> (2) Very large result s
Hi,
I will push the commit shortly.
Regards,
Thomas
On Wednesday, April 15, 2015, bocher wrote:
>
> Hi,
>
> Please read https://github.com/h2database/h2database/pull/2.
>
> Erwan
>
>
> Le mardi 14 avril 2015 19:31:09 UTC+2, wener a écrit :
>>
>> Any explain of "EXTERNAL" table type ? Glad to
ke SQL queries from any standard industry reporting
> tool like Birt, jasperReports, etc.
>
> I've provided a bit large answer because I think this is an interesting
> business case for H2 as it can be used to wrap any custom language (even
> Lucene queries, for instance) into stand
source.html?file=org/h2/mvstore/db/MVSpatialIndex.java&line=103&build=186>)
>
> at org.h2.mvstore.db.MVTable.addIndex(MVTable.java:433
> <http://h2database.com/html/source.html?file=org/h2/mvstore/db/MVTable.java&line=433&build=186>)
>
> at org.h2.com
Hi,
There was a problem migrating the Google Code issues to Github. Maybe this
can still be resolved, but maybe its simpler to just keep the old issues in
Google Code and not migrate them.
For now, I would prefer if you just send an email with the issue report to
this group instead of using the i
Hi,
> how to get the where clause?
Well, why do you need it exactly?
Regards,
Thomas
On Sun, Apr 12, 2015 at 6:39 PM, Pablo Beltran > wrote:
> Hi Fred,
>
> I've searched for the condition attribute in the org.h2.command.dml.Select
> class source code and it's declared as private:
>
> *private
Hi,
> If this were just my local database I would do upgrade myself, but since
this "old" version is on customer side, and most of them are computer
unsavy this would be a no go...
Well, you could automate that...
Regards,
Thomas
On Tue, Apr 14, 2015 at 9:30 AM, Andy Rozman > wrote:
> Ok. So
my purpose.
>
> Any suggestion?
>
> Regards,
> Arun
>
>
> On Friday, April 10, 2015 at 9:33:04 PM UTC+5:30, Thomas Mueller wrote:
>>
>> Hi,
>>
>> You would need to use a more recent version of H2. The version you are
>> using does not support serializ
Hi,
As I already wrote, it depends on what you do exactly. If you don't give us
that info, then we can't help I'm afraid.
Regards,
Thomas
On Sun, Apr 12, 2015 at 12:55 PM, Cecil Westerhof > wrote:
> 2015-04-12 12:03 GMT+02:00 Fred&Dani&Pandora&Aquiles >:
>
>> Hi Cecil,
>>
>> I tried that, but
Hi,
Yes, I added a timeout.
First I wanted to keep the CLOB / BLOB objects only until the next
transaction. The problem is that some methods in JdbcConnection started a
new transaction (Statement.getQueryTimeout for example). It would be better
if those transactions are not required, but thats a
Hi,
Thanks! My mistake, I didn't copy all the files... I fixed it now, but it
will take some more time until it is synchronized.
Regards,
Thomas
On Sat, Apr 11, 2015 at 4:29 PM, bocher wrote:
>
> Dear Thomas,
>
> The maven repository contains only the sources for the last release
> 1.4.187 (h
Hi
The Google Group will be kept.
Regards,
Thomas
On Friday, April 10, 2015, Fred&Dani&Pandora&Aquiles
wrote:
> Hi,
>
> The discussions about the requests will move there now too, or you'll
> still keep this group for a time?
>
> Regards,
>
> Fr
Hi,
The source code of H2 is now migrated to Github:
https://github.com/h2database/h2database
The list of issues could not be migrated, I hope I can still do that, but
if not I guess it's not a big issue.
The links from the web page will need to be changed.
Regards,
Thomas
--
You received thi
Hi,
As I already wrote, the bugfix and a test case are already committed in the
trunk.
Regards,
Thomas
On Thu, Apr 9, 2015 at 6:15 PM, John McMahon
wrote:
> Hi Thomas,
>
> Did you have time to look into this?
>
> Thanks
> John
>
> On 8 April 2015 at 09:18, John McMahon wrote:
>
>> Minor corr
Hi,
I found some more issues with nioMemLZF, I will fix that as well. And add a
test case. Also, the page sizes are probably too small, I will check that.
By the way the prefixes are "nioMemFS:" and "nioMemLZF:".
Regards,
Thomas
On Thursday, April 9, 2015, Ravi Sanwal > wrote:
> There is a b
Hi,
I think you are missing an index on table B, column A.
But probably the explain plan is a big confusing as well, not sure how to
be fix this.
Regards,
Thomas
On Tuesday, March 17, 2015, Fred&Dani&Pandora&Aquiles
wrote:
> Hi,
>
>
> There are 2 tablescans in the query, there should only
Hi,
Sorry I don't understand the question. After shutdown, the database is
closed of course.
Regards,
Thomas
On Thursday, April 9, 2015, sim wrote:
> Hi,
>
> db returns in multi_threaded false state after executing shutdown defrag.
>
> Is that a normal behaviour?
>
>
> --
> You received thi
Hi,
I would say, it depends on what you do exactly...
Regards,
Thomas
On Thu, Apr 9, 2015 at 12:13 PM, Cecil Westerhof > wrote:
> The first time I send a query to H2 it takes around 25 seconds. All
> subsequent queries are ready in less as a second. Is this normal? Can I
> circumvent this?
>
>
Hi,
We added db_close_delay because we have tests that open a Hibernate Session
> (connection), do some work, close that Session (thus closing that
> connection), open a new Session and try to do some work with the stuff done
> in the initial Session. But H2 had dropped the database in between.
>
RUE
>
> I had to change that everywhere to instead
> use jdbc:h2:mem:db1;DB_CLOSE_DELAY=-1
>
> On Wed, Apr 8, 2015 at 7:37 AM, Thomas Mueller <
> thomas.tom.muel...@gmail.com
> > wrote:
>
>> Hi,
>>
>> What database URLs do you currently use?
>>
>
Hi,
What database URLs do you currently use?
Maybe we should add a way to auto-delete the database file if it is closed?
This could be done in the file system abstraction, for example using the
prefix "autoDelete:", so that a database URL would look like
"jdbc:h2:autoDelete:./target/db/test" (usi
Hi,
Where is it documented ?
>
That's a good point. It is not well documented currently, I will fix that.
The documented limitations only apply to the PageStore storage engine,
which is the default in version 1.3.x. With version 1.4.x, the MVCC mode is
enabled by default, and better tested, and a
Hi,
I would probably try version 1.3.176 (2014-04-05), which is marked as the
"last stable" version.
The method DatabaseMetaData#getSchema(String,String) is there since about
February 2011, but threw an exception back then ("unsupported"). In
February 2013 (revision 4659) the method was correctly
Hi,
Yes, use
SELECT * FROM INFORMATION_SCHEMA.LOCKS;
SELECT * FROM INFORMATION_SCHEMA.SESSIONS;
To cancel a transaction of a different connection, see
http://h2database.com/html/functions.html#cancel_session
Regards,
Thomas
On Tue, Apr 7, 2015 at 12:20 PM, Антон Чехов > wrote:
> Hi Thomas Th
Hi,
Yes, this issue was reported last week as well; it should be fixed in trunk.
Regards,
Thomas
On Tue, Apr 7, 2015 at 12:11 PM, John McMahon > wrote:
> We are experiencing null pointer exceptions when we attempt to store a
> null geometry in H2 when we close the connection
>
> Attached is a s
Hi,
Yes, this issue was reported (issue 605) and is now fixed in the trunk. I
wanted to release a new version this weekend, but unfortunately I was sick.
Right now I'm trying to fix an important issue (also related to LOBs);
hopefully there release will be this weekend.
Regards,
Thomas
On Friday
Hi,
I can't reproduce this problem. Could you post a simple test case please?
Regards,
Thomas
On Tuesday, April 7, 2015, Serge Korn wrote:
> Hi,
>
> Since upgrading to the latest beta release I noticed that "order by" is
> ignored when used in combination with "where" clause.
> When I downgrad
Hi,
Descending iteration is not yet implemented.
If performance is not the most important part, it could be emulated easily
(with a _lot_ less code) using MVMap.getKeyIndex(K) and getKey(long).
Possibly you also need MVMap.openVersion. That way, iterating over n
entries would then be O(n log n) i
Hi,
This error message means the table is locked by another connection.
Increasing the timeout will only help if the other connection commits the
transaction in the meantime.
Regards,
Thomas
On Monday, April 6, 2015, Антон Чехов wrote:
> Hi All,
>
> Could you please help.
> When i login to h2
Hi,
> Why not extract the mvstore, make H2 more flexible.
>
It is already, and there is a separate jar file for it.
Regards,
Thomas
--
You received this message because you are subscribed to the Google Groups "H2
Database" group.
To unsubscribe from this group and stop receiving emails from i
Hi,
H2 supports the standard CSV format as documented in Wikipedia, including
RFC 4180, and much more. If MariaDB doesn't support that (are you sure it
doesn't?), then I suggest to ask _them_ why they are not compliant. You
should probably ask at StackOverflow.com.
But yes, JSON export and import
3, 2015 at 3:36:55 PM UTC+3, Thomas Mueller wrote:
>>
>> Hi,
>>
>> You may have read already that Google Code will shut down. That means the
>> source code of H2 will be moved, most likely to GitHub. This move will
>> happen shortly, in the next weeks or months
Hi,
Thanks a lot for reporting this issue! This should be fixed no (in the
trunk).
Regards,
Thomas
On Wed, Mar 18, 2015 at 3:14 PM, Nicolas Fortin (OrbisGIS) <
nico.de...@gmail.com> wrote:
> Hi,
>
> Sorry for double post. Mail client issues.
>
> Regards,
>
> Nicolas Fortin
>
> Le vendredi 13 ma
t now, could anybody test it? If not, I
will install the latest version and test later on.
Regards,
Thomas
On Thursday, March 26, 2015, Noel Grandin wrote:
>
>
> On 2015-03-26 08:48 AM, Thomas Mueller wrote:
>
>>
>>
>> I didn't know that. I made some tests (se
document ?
>
> 在 2015年3月10日星期二 UTC+8上午1:15:00,Thomas Mueller写道:
>>
>> Hi,
>>
>> You would need to use the server mode, or the automatic mixed mode.
>>
>> Regards,
>> Thomas
>>
>>
>> On Saturday, January 31, 2015, Peter Cheung wrote:
Hi,
(1) add a manual command to clean them up
>>
>
Yes, that would be an option, and we would not have to change behavior by
default (which is a risk as it could break existing applications).
> (2) change it so that the client can only read from a BLOB while a
>> transaction is open. Flush any t
n the tables?
> Is there an advantage to one over the other?
>
> On Tuesday, March 24, 2015 at 2:42:23 AM UTC-4, Thomas Mueller wrote:
>>
>> Hi,
>>
>> Sorry I only tested with the latest version of H2. I think in previous
>> versions, paths were not supported for
Hi,
It works for me. Maybe you use a 1.4.186 client, but an old server version?
Regards,
Thomas
On Tue, Mar 17, 2015 at 2:09 AM, Andy Glick > wrote:
> When we submit the following SQL query to either of H2 database revisions
> 1.4.180 or 1.4.186 the query executes successfully:
>
> select time
Hi,
Such questions are best asked at StackOverflow.com.
I think the best solution would be to use a spatial index.
Regards,
Thomas
On Friday, March 20, 2015, Uli wrote:
> I had the same question but no good answer yet:
> https://groups.google.com/forum/#!topic/h2-database/HtgVV1pglao
>
> --
Hi,
Thanks a lot for the great test case! This is a bug that was introduced in
the last version, in a mechanism to remove unreferenced CLOB and BLOB
blocks. This mechanism sometimes removes entries that are still used. I
have a fix now and will commit it to trunk.
Regards,
Thomas
On Monday, Mar
Hi,
> Both the server and the client are using H2 1.4.186.
No, that's not the case. You reported the error message is "Column "TS" not
found; SQL statement: select TS '2015-03-16 12:12:12.12' [42122-170]". The
"170" at the end of the error code means you are using version 1.3.170 on
the server si
Hi,
Sorry I only tested with the latest version of H2. I think in previous
versions, paths were not supported for in-memory databases. That means, you
would need to use a simple name, for example "jdbc:h2:mem:gpfp" or similar.
Regards,
Thomas
On Wednesday, March 18, 2015, Jason wrote:
> The da
Hi,
H2 does have a "cluster" feature, but it is quite limited, see the
documentation.
But I would probably try out https://github.com/shesse/h2ha
Regards,
Thomas
On Tue, Mar 17, 2015 at 6:23 PM, Martin Kersten > wrote:
> Hello H2 Group,
>
> I use H2 since my early Java days (around 2000 I
Hi,
I would consider using a spatial index.
I think no matter what you do, regular b-tree indexes can't be efficient if
you have many (millions of) entries: no matter what index you pick, some
queries would still scan about 50% of all entries.
Maybe you want to ask this question at StackOverflow
he the value (Huge value) for a specific
> time until the one process is complete. I used LinkedHashMap at first but
> unfortunately it couldn't hold the big values.
> Am attaching the code here. Please have a look and let me know if any fine
> tuning has to be done
>
> Re
Hi,
Changing the collation: you would need to export / import all data. Please
note a collation is slightly slower than ignore-case.
> select cast(name as VARCHAR_IGNORECASE) from TEST_SORT() order by 1
Yes, that's a good solution / workaround. I would probably use that if
possible.
Regards,
Th
Hi,
> INFORMATION_SCHEMA.lobs table grows rapidly
Yes. This is a problem for both MVStore and PageStore. The problem is that
CLOB and BLOB objects must still be available even after the transaction is
complete, and I think even after the result set is closed. And even after
the row is removed. In
Hi,
There is no case-insensitive SQL type, so changing SimpleResultSet might be
tricky / might not work.
I did not test this, but would this work for you (without patch):
SET COLLATION ENGLISH SECONDARY
Regards,
Thomas
On Mon, Mar 23, 2015 at 9:26 AM, wrote:
> Hi Fred,
>
> Thanks for your p
March 19, 2015, Eric Chatellier
wrote:
> Le 19/03/2015 07:45, Thomas Mueller a écrit :
> > Hi,
> >
> > The settings can be changed in any order... Does it consistently work in
> one
> > case, but not the other?
> Yes, it consistently fails if "set clust
Hi,
I have a fix and will commit it in trunk. A workaround is to enable
autocommit for the "runscript" part.
Regards,
Thomas
On Thursday, March 19, 2015, Eric Chatellier
wrote:
> Le 19/03/2015 07:45, Thomas Mueller a écrit :
> > Hi,
> >
> > The settings can be
Hi,
The settings can be changed in any order... Does it consistently work in
one case, but not the other?
Could you try with the latest version of H2, and append ";mv_store=false"
to the database URL? I can't reproduce the problem with the latest version,
and I didn't try yet with the version you
?
>
>
> On Tuesday, March 17, 2015 at 7:05:57 PM UTC+2, Thomas Mueller wrote:
>>
>> Hi,
>>
>> Yes, there is the "retention_time" setting. Did you try that already? I
>> would be interested how much this helps in your case.
>>
>> Regards,
&g
Hi,
Yes, there is the "retention_time" setting. Did you try that already? I
would be interested how much this helps in your case.
Regards,
Thomas
On Sunday, March 15, 2015, sim wrote:
> Hi,
>
> I have been testing a real life application where use H2 1.4.186 with
> mvstore on. The applicatio
Hi,
This is not directly supported. You could write a wrapper for that, which
internally uses two maps. If your map is of type (K => V), then the first
map would be (K -> (V, counter)), and the second map would be (counter =>
K). That way, you can do efficient lookup by the counter and the key.
B
Hi,
Could you find out what the final database URL is? It looks like some kind
of formatting problem, but I'm not sure.
Regards,
Thomas
On Fri, Mar 13, 2015 at 2:43 PM, Jason > wrote:
> Sorry, here's the entire trace. I'm using Mybatis for object-relational
> mapping.
>
> Exception in thread
t;
> --
>
> Nicolas Fortin
> IRSTV FR CNRS 2488
> GIShttp://orbisgis.org
> Spatial DB http://h2gis.org
> Noise http://noisemap.orbisgis.org
>
> Le jeudi 12 mars 2015 18:50:11 UTC+1, Thomas Mueller a écrit :
>>
>> Hi,
>>
>> At some poi
Hi,
Strange, it only seems to be a problem if there is an "alter table". If X4
is added at the beginning, then it's fine.
Regards,
Thomas
On Friday, March 13, 2015, Nicolas Nico wrote:
> Hi,
>
> With H2 1.4.186 we are experimenting a performance decrease between
> PageStore and MVStore.
>
>
Hi,
Correction: the URL will probably be http://github.com/h2database/h2
I just created http://github.com/h2database
Regards,
Thomas
On Fri, Mar 13, 2015 at 1:46 PM, Sergi Vladykin
wrote:
> Sweet!
>
> Sergi
>
>
> 2015-03-13 15:36 GMT+03:00 Thomas Mueller :
>
>>
401 - 500 of 5365 matches
Mail list logo