Hi,
C. SELECT pk FROM t WHERE pk = ? OR pk = ? [7688ms]
Nearly 8 seconds, is that right?
Yes. This statement is not optimized at all, a table table scan is
made. It is planned to improve that, see Optimize ID=? OR ID=?:
convert to IN(...) on the roadmap. However the priority of this item
is
Hi,
13:06:49.767: Recovering C:\DataCache\h2\testDB.48.log.db 0%
13:06:53.485: Recovering C:\DataCache\h2\testDB.48.log.db 1%
It looks like the transaction log file is quite large. How large are
the .log.db files? The maximum size can be set using SET MAX_LOG_SIZE,
see also
Hi,
there is exception when I
try to persist a CLOB field.
What exception did you get? The database exception should be in the
.trace.db file, but the Hibernate exception would be helpful as well.
Could you post a simple test case (one table, one class with a main
method) so I can reproduce
Hi,
Thank for your help! It looks like this is a bug in the transaction
log file size calculation at Database.setMaxLogSize(). The log file
should not be that large. I will fix this for the next release.
Unfortunately, it looks like there is no workaround at the moment
(except changing
Hi,
What database URL do you use? How many tables and other database
objects (triggers, sequences, columns,...) do you use? From the memory
dump the following line is interesting: org.h2.index.TreeNode
107,219,552 bytes. The class TreeNode is used for in-memory indexes.
The in-memory index is
Hi,
Thank you for adding this method. It is, however, throwing an
exception back at me.
Strange... What database URL do you use?
Regards,
Thomas
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups H2
Database
Hi,
jdbc:h2:/tmp/prod/
h2cache;TRACE_LEVEL_FILE=3;CACHE_SIZE=1048576;LOCK_MODE=3;MVCC=TRUE;MULTI_THREADED=0
This means MULTI_THREADED is disabled in your case.
But I am opening and using several Sessions, concurrently, from
different threads .
This shouldn't be a problem.
I believe, for
Hi,
Instead of
st.executeUpdate(CALL FTL_CREATE_INDEX('PUBLIC', 'test_table_li',
NULL););
try
st.executeUpdate(CALL FTL_CREATE_INDEX('PUBLIC', 'TEST_TABLE_LI',
NULL););
Regards,
Thomas
--~--~-~--~~~---~--~~
You received this message because you are
Hi,
You have 3 options:
- You give the database enough rights to create the directory.
- You create the directory before starting the application.
- You use a directory where the application has enough access rights.
on Windows the directory /ourb is being created
Because the process has the
.engine.SessionRemote.createTempFile(SessionRemote.java:457)
... 8 more
On Oct 1, 9:34 pm, Thomas Mueller [EMAIL PROTECTED]
wrote:
Hi,
I can reproduce this problem with previous versions, but so far I
couldn't reproduce it with version 1.0.78. Could you post a simple
test
:buildOK
%JAVA_HOME%/bin/java -Xmx512m -cp
bin;%JAVA_HOME%/lib/tools.jar;temp org.h2.build.Build %*
Regards,
Thomas
On Sun, Oct 5, 2008 at 9:43 PM, Thomas Mueller
[EMAIL PROTECTED] wrote:
Hi,
I tried to build JaQu per the instructions ( which are wrong for unix/
OS X as the script is build.sh
Hi,
In my application I have my own USERS table and this seems to conflict
with the system USERS table.
When I try to get meta data via jdbc, I get a combo of my table and
the system table.
resultSet = metaData.getColumns(null, , tableName, %);
while (resultSet.next()) {
final
Hi,
Of course you could just use the server mode.
I believe there is a new connection mode that will create and manage
the server for you. But I'm not sure how to do that. So I second your
question: How do we allow multiple connections to embedded database?
Yes there is a new (experimental)
Hi,
In Oracle, this will throw an error. Any chance H2 can do the same?
It doesn't throw an exception in PostgreSQL. If it would throw an
exception in H2, then H2 wouldn't be compatible with PostgreSQL.
alter table user_ref drop CONSTRAINT ui_ur_fk;
drop table user_info;
drop table user_ref;
Hi,
Thanks for your mail! This is a bug in the optimization to group using
an index. This feature is there since a long time, but was always
disabled by default. In version 1.1 (which is still beta) it is
enabled by default. You can disable this feature by setting the system
property
Hi,
Thanks for the patches! I have changed it, the newest version is
committed to the trunk at:
http://code.google.com/p/h2database/source/browse/trunk/h2/src/main/org/h2/table/TableLink.java
Regards,
Thomas
On Fri, Oct 10, 2008 at 7:52 PM, Pavel [EMAIL PROTECTED] wrote:
After the fix I also
Hi,
I believe that it should return 0.
The JDBC API says that executeUpdate should return:
either the row count for INSERT, UPDATE or DELETE statements, or 0
for SQL statements that return nothing
Other databases also return 0.
Regards,
Thomas
On Fri, Oct 10, 2008 at 10:30 PM, Pavel [EMAIL
Hi,
Thank for your help! I will have a look how this can be improved.
Maybe it makes sense to use int or byte, but I want to make sure it
does in fact get faster so I will need to write some benchmarks as
well.
Regards,
Thomas
On Sat, Oct 11, 2008 at 12:00 AM, Alex [EMAIL PROTECTED] wrote:
) the file IO came up and became the slowest operation.
Alex
On Oct 13, 3:19 pm, Thomas Mueller [EMAIL PROTECTED]
wrote:
Hi,
Thank for your help! I will have a look how this can be improved.
Maybe it makes sense to use int or byte, but I want to make sure it
does in fact get faster so I will need
Hi,
The code synchronizes on a new object each time - that's like no
synchronization at all. You probably want: synchronized
(ContextInfo.class) { ... }
What I'm trying to do is get a connection at the beginning of the
request and return it to the pool at the end of the request.
The rest of
Hi,
I am sorry to say that, but it looks like a corruption problem. I am
very interested in analyzing and solving this problem. Corruption
problems have top priority for me. I have a few question:
- Could you send the full stack trace of the exception including message text?
- What is your
Hi,
SET AUTO_SERVER = mixed mode?
No. 'Mixed mode' can mean two things:
(A) start a server manually and then connect to it.
(B) AUTO_SERVER=TRUE in which case the server is started automatically.
Regards,
Thomas
--~--~-~--~~~---~--~~
You received this message
Hi,
Thanks a lot! I have now integrated the message. This will be included
in the next release.
Regards,
Thomas
On Sun, Oct 12, 2008 at 12:53 AM, Dario Fassi [EMAIL PROTECTED] wrote:
Hi all,
I have done a acceptable (not professional) translation of
messages_en.properties to Spanish (es).
Hi,
insert into user_info(user_info_id) values ('9');
insert into user_info(user_info_id) values('09');
insert into user_ref(user_ref_id) values (009);
The user_ref entry now points to which user_info record?
For PostgreSQL, the first one ('9').
If select 'x' from dual caused postgres to
Hi,
Could you post the data or a test case with random data?
Regards,
Thomas
On Thu, Oct 16, 2008 at 5:01 AM, Abyric [EMAIL PROTECTED] wrote:
After thouroughly reading trough many SQL optimisation sites I still
fail to see what I can to to increase performance. To keep it short,
my query
Hi,
Thanks for the test case! The optimizer took most of the time, it was
calculating the cost of too many query plans. I have now fixed this
problem, now the query takes less than 100 ms for me (the first time
is a bit slower).
Regards,
Thomas
Hi,
1)Auto_server provides a new access method to embedded mode since in
old embedded mode only one process could possible access the db.
2)Mixed mode means starting a server listening on a port along with a
embedded mode enable, while remote access to the db is
possible.Meanwhile,the
Hi,
Thanks for your mail!
when a (b/c)lob record is deleted, the
file is renamed from 1.t75.lob to 1.temp.lob, and then the file may,
or may not, be deleted when the database is shutdown.
Yes, the file is renamed to *.temp.db when the transaction is committed.
I couldn't find any
Hi,
Thanks! Your solution is faster, but I like to wait implementing it
unless it is really a measurable performance problem, because the
solution is quite complex :-) Maybe I will just switch to BitSet at
some point.
Regards,
Thomas
On Thu, Oct 16, 2008 at 5:25 PM, Alex [EMAIL PROTECTED]
Hi,
select distinct a0.word from contexts.words a0 where (a0.word =
cast('apple' as varchar)) and ((select count(a2) from (select a0.word
as a2) a3) = (select 1)) order by a0.word
Sorry I don't understand what the query should do. Could you explain
it please? Is it the same as:
select word
Hi,
Is a clustered index supported?
It depends on what exactly you mean with clustered index... Could you
explain what you need and why?
Regards,
Thomas
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups H2
Hi,
TEXT is currently mapped to CLOB. Unfortunately I don't have a
workaround for you, I'm sorry. I will add a feature request to allow
to remap such data types, but it will take some time until this is
implemented.
Regards,
Thomas
On Mon, Oct 20, 2008 at 10:06 PM, [EMAIL PROTECTED]
[EMAIL
Hi,
Thanks for your mail! So far I could not reproduce the problem. Could
you (that would be idea) create a simple test case? If not, could you
run the test with the trace settings, using a fresh database? Just
append ;TRACE_LEVEL_FILE=3 to the database URL, and then post the
.trace.db file.
Hi,
From what I understand, a non-clustered index will store a key-value
mapping outside of the database where the keys are ordered and the
values are locations of records. A clustered index will actually
store records in sorted order within the database itself, or so I
think.
There is a
.
Regards,
Thomas
On Oct 23, 8:06 am, Thomas Mueller [EMAIL PROTECTED]
wrote:
Hi,
A solution is to use
CREATE TABLE Contents(c1 SMALLINT, id SMALLINT, Bid SMALLINT, Seq SMALLINT,
off
VARCHAR(10), Duration VARCHAR(10), Line VARCHAR(1))
AS SELECT * FROM CSVREAD...
Regards,
Thomas
On Wed
Hello,
A new version of H2 is available for download at http://www.h2database.com
(you may have to click 'refresh').
Changes and new functionality:
Changes and new functionality:
- The French translation of the H2 Console has been improved by
Olivier Parent.
- Translating the H2 Console is
Hi,
I've taken a look at the flushRedoLog method in the DiskFile. I may be
wrong, but could you explain how it supposed to work.
It is first writing all deleted blocks, and then the blocks that are
not deleted. However for a given position, only the last entry is
written.
If my log
Hi,
we noticed that the log flush is very inefficient.
it writes out every block of 128 bytes as a separate file
write. On slow files systems (flash) this causes a ton of unnecessary
overhead.
I never tested flash file systems. Based on your description, regular
database operations are
Hi,
it is possible to set a trigger on a view.
So it could be possible to make a changable view with a trigger.
But in the class org.h2.table.TableView an exception in the methods
addRow and removeRow is thrown.
It should be possible to make the throwing von exceptions in this
methods
Hi,
I cannot find anything about bitmap index in the documentation. I
suppose its not implemented am I right?
Correct, bitmap indexes are not implemented.
What are the best practices for columns with small amount of distinct
values?
There are no special best practices.
How H2 treats
Hi,
Once my db shards grow beyond 500,000 records
Sorry, what is a 'db shard'?
Inserting 5 million records over 64 clean db shards happens in
3 minutes.
So you mean 64 databases?
After those shards get over 500,000 records - the whole
insert takes 29 minutes.
Maybe the system runs out
Hi,
I think I understand the problem now a bit better. I have a simple
test case that also fails on H2, but works on PostgreSQL, MySQL and
Apache Derby:
create table a(id int);
create table b(id int);
create table c(id int);
select * from a left outer join (b left outer join c on b.id = c.id)
Hi,
All of our unit tests are really functional tests that actually exercise
Postgres.
So your test cases work with PostgreSQL, but don't work with H2?
setProperty(hibernate.dialect, org.hibernate.dialect.H2Dialect).
Did you try using the PostgreSQL dialect for H2? That should work -
Hi,
This is an incompatibility with PostgreSQL. I will fix that in the
next release. A workaround is to use:
select nextval('CONF', 'REPORT_ID_SEQ');
Unfortunately this is not supported by PostgreSQL.
Regards,
Thomas
On Sat, Oct 25, 2008 at 4:24 AM, Limbic System [EMAIL PROTECTED] wrote:
Hi,
This does not seem to be the case though...
You are right, thanks for the test case! Obviously I forgot why the
code is like that. I should have documented it better... I think I
remember now why all delete entries need to be written first:
Because delete entries are always 1 block, while
Hi,
At the moment I'm doing more
thorough benchmarks that I've created.
If you are running benchmarks, could you benchmark them please? To do
that use java -Xrunhprof:cpu=samples,depth=8 ... - and then
upload/post the file java.hprof.txt
If the benchmark application is not a self contained
Hi,
Hi. I am using Win XP and my H2 version is 1.1.100. I tried to 'cd' to
the \h2 directory in my command prompt and type in 'build jarSmall'
and it showed Error: JAVA_HOME is not defined. 'javac' is not
recognized as an internal or external command, operable program or
batch file.
You
Hi,
Sorry this is not possible at the moment. There is a feature request
already, but it will take some time until this is available. A
workaround is:
select xm, sum from(
select '' o, '总计' xm, sum(je) sum from #t union all
select col1 o, col1, sum(je) from #t group by col1 union all
select col1
Hi,
It works for me. Could you run this and post the result?
System.getProperties().list(System.out);
Thanks,
Thomas
On Wed, Oct 29, 2008 at 6:01 PM, Piercio [EMAIL PROTECTED] wrote:
I am using H2 1.1.102 in embedded mode, and could not insert the date
2042-10-12 in the database.
Using
Hi,
Thanks a lot for the test case! I am working on a fix for this problem
now. I believe this will be fixed in the next release. The problem is
that for MVCC the rows are copied in one large transaction (CREATE
TABLE... and then INSERT INTO .. SELECT). I will try to use CREATE
TABLE ... AS
Hi,
Eclipse doesn't need to 'handle' them. Those are regular Java block
comments. By default the source code uses Java 1.6 features, however
Java 1.4 and 1.5 are supported as well. To disable Java 1.6 specific
features, run:
build switchSource
See also http://www.h2database.com/html/build.html
much.
On Oct 29, 4:10 pm, Thomas Mueller [EMAIL PROTECTED]
wrote:
Hi,
It works for me. Could you run this and post the result?
System.getProperties().list(System.out);
Thanks,
Thomas
On Wed, Oct 29, 2008 at 6:01 PM, Piercio [EMAIL PROTECTED] wrote:
I am using H2 1.1.102 in embedded mode
Hi,
FullText.createIndex(conn, PUBLIC, TEST, null);
FullText.dropIndex(conn, PUBLIC, test);
The table name is case sensitive like the DatabaseMetaData methods,
that means it must match the table name as it is stored in the
database. So you need to write:
Hi,
I was wondering is it possible to use H2 to parse sql and get some
kind of abstract syntax tree. I have looked and it does not appear to
be easy to use in this way. Can someone confirm this?
It depends what you need. When parsing (class Parser), H2 constructs
objects of type 'Prepared'
Hi,
Would that be possible to make a change so that MAX_LOG_SIZE can also
be specified in kilobytes? Something like, if the value ends in 'K',
it means kilobytes, Nothing or 'M' at the end means megabytes.
I would rather solve the performance problem.
SET MAX_LOG_SIZE=1 // setMaxLogSize(1
Hi,
PreparedStatement pstmt = connection.prepareStatement();
...
ResultSet rs = pstmt.executeQuery();
while(rs)
{
a lot of small request generated from rs content
}
Most likely the problem is that the result set is closed while
processing the nested requests. Currently H2 uses a
Hi,
my scenario: web application using H2 for demo purposes.
I'm not sure, did you already read the documentation at
http://www.h2database.com/html/tutorial.html#web_applications
We'd like to embed the H2 Console into the existing web application,
so it's reachable on the same
Hi,
What is the best way to show H2 content in jTable?
I don't know JTable. I suggest to use Google and search for jtable
jdbc or (if you use Hibernate) jtable hibernate.
getValueAt - SELECT * FROM table1 WHERE id = ?
My idea is to cache ResultSet in case next getValueAt would point to
because it's slow, complicated, and seems to be buggy.
However it will take some time until this is done, and it can't solve
problems with java.sql.Date / Time / Timestamp.
Regards,
Thomas
On Fri, Oct 31, 2008 at 4:19 PM, Thomas Mueller
[EMAIL PROTECTED] wrote:
Hi,
FYI I am discussing
, so they
ought to be GCd (or so I think).
-Julian
On Oct 29, 1:50 pm, Thomas Mueller [EMAIL PROTECTED]
wrote:
Hi,
but as the tables get beyond 500,000 rows ...
the performance gets brutally slow.
I can't reproduce this so far. Could it be a cache / memory problem?
If you
Hi,
I have a question about the embedded mode:
How many applications can read/write the db files at same time?
You can have multiple connections. By default access to the database
is serialized however. To disable this, call SET MULTI_THREADED 1 -
see
Hi,
Yes, we use LOG=2
It obviously helps, but when the log file gets big, it still takes
several minutes to work through it.
What version of H2 do you use? Could you profile the application to
find out why it takes so long? To do that, write a simple application
that only connects to the
Hi,
Could you try connecting to the server using localhost? Just replace
the ip address with 'localhost'. For me, the default URL is:
http://192.168.0.100:8182
but this also works:
http://localhost:8182
If this solves the problem then maybe it is a firewall / virus scanner
problem. Could you
Hi,
I don't know Spring getJdbcTemplate. Maybe Spring uses
ResultSetMetaData.getColumnName instead of getColumnLabel. There is a
way to change the behavior of H2: you need to set the system property
h2.aliasColumnName to 'true', see also
Hello,
A new version of H2 is available for download at http://www.h2database.com
(you may have to click 'refresh').
Changes and new functionality:
- New column INFORMATION_SCHEMA.COLUMNS.SEQUENCE_NAME.
- Aliases for built-in data types can now be re-mapped.
- Improved PostgreSQL compatibility
Hi,
When I first load up the console shouldn't the JDBC URL be defaulted
to the one in on the context-param? Or do I have it in the wrong
place?
The default database URL of the console is stored in the user
preferences, for Windows, that is:
C:\Documents and
Hi,
Thanks for your help! I have voted for the bug.
Regards,
Thomas
On Sat, Nov 8, 2008 at 8:44 PM, btyla [EMAIL PROTECTED] wrote:
Hi,
EclipseLink is Oracle's TopLink code donated to the Eclipse foundation
to build the JPA 2.0 reference implementation. EclipseLink is
currently lacking
Hi,
Is there a way to truncate a table with cascade, so that all rows in
other tables which reference the table being deleted are also deleted?
Sorry, there is now way to do that (that I know of). I will add a
feature request for Support TRUNCATE .. CASCADE like PostgreSQL, but
unless other
Hi,
the code of the H2 Platform class suggested in the bugzilla issue
above fails with CLOB/BLOB fields because of the issue we discussed
earlier.
This problem will be solved in the next release. Details:
ResultSet.getObject for CLOB or BLOB will return a java.sql.Clob /
java.sql.Blob object
Hi,
I got a database with one open connection. If I call close() will the
database be closed for good after the call returns?
By default, yes, the database is closed when the last connection is
closed. You can prevent that using the DB_CLOSE_DELAY, see also
Hi,
Is there a way to override the default? I want to be able to ship a
custom properties file that automatically has the JDBC URL set and not
have it get created and set in that temp folder.
You could just include a 'good' properties file as a resource in your
application and overwrite the
Hi,
I plan to use H2 for a very small database (two tables with column
pair key-value) The default cache size is 20 MB and thats too much
because I am having two embedded databases opened in one java process.
Thats 20+20 = 40 MB of cache.
What is recommended cache value for such a small
Hi,
The problem for AVG will be fixed in the next release. There is a
problem for SUM as well: it can overflow when using large values. In
the next release, SUM will return a data type that is safe. I hope
this will not break applications that call ResultSet.getObject and
expect the old data
. If it does not, you could change the system property, or
(even better) store the result of FTL_SEARCH in a temporary table
first.
Regards,
Thomas
On Mon, Nov 17, 2008 at 7:21 PM, Thomas Mueller
[EMAIL PROTECTED] wrote:
Hi,
Could you run
SCRIPT NODATA TO 'test.sql';
and then send me this file
Hi,
I am sorry to say that, but it looks like a corruption problem. I am
very interested in analyzing and solving this problem. Corruption
problems have top priority for me. I have a few question:
- Could you send the full stack trace of the exception including message text?
- What is your
Hi,
Multi version concurrency is specially made to avoid blocking readers.
But it looks like you want to explicitly lock readers. So you can't
use Multi-version concurrency. In H2, what you could do is:
Open two connections. On both connections, call:
SET LOCK_TIMEOUT 60 * 1000 (that is one
, Thomas Mueller
[EMAIL PROTECTED] wrote:
Hi,
I am not working with hibernate so I won't be much help with that, but
I am all for completing/improving the parser bit of it. The source
archive is only about 50kb, I won't email it to the list. Would you
like me to email it to you perhaps
Hi,
We are using the H2 database for creating permanent tables that
contain huge data sets. This table gets deleted automatically after a
certain time interval.
We have our own cache clearing logic that works
every hour on H2 but the tables get deleted automatically by the H2
caching
Hi,
I am trying to replace a PostgreSQL database with H2.
In that case, you need to use the H2 JDBC driver (org.h2.Driver
instead of org.postgresql.Driver), and the H2 JDBC URL (jdbc:h2:...
instead of jdbc:postgresql:...).
First, I connected normally :
Class.forName(org.h2.Driver);
Hi,
I didn't know about generate_series, maybe I will support it in future
versions. H2 supports this:
SELECT X FROM SYSTEM_RANGE(1, 100);
Regards,
Thomas
On Mon, Nov 24, 2008 at 8:51 PM, Marek Kubis [EMAIL PROTECTED] wrote:
Hi
I'd like to generate integer number series in my queries
Hi,
If you create a database with 100 or more tables, the tree menu which lists
all the tables doesn't seem to work properly.
That's not a bug it's a feature. For some databases reading the
schema is really slow, and I wanted to avoid that when there are many
tables. However for H2 it's not
Hi,
The stacktrace is :
org.h2.jdbc.JdbcSQLException: Connection is broken [90067-100]
When I connect Squirrel SQL, I use the absolute path like
jdbc:h2:tcp://localhost:9092/data/servers/glassfish/domains/myApp/
config/db
If you use the server mode, you need to start a server. I'm not sure
Hi,
Is there a way to specify how Lucene should index (e.g. multiple
fields / different analyzer per field (based on the columns)) and
query (e.g. different weight on each field) the columns with full-text
search? This doesn't need to be SQL solution, I'm fine if there is a
Java API to do
Hi,
select id, convert(text,int) as number from test where type='int'
This will first check the condition and then convert only the rows
where it's not a problem.
select * from (select id, convert(text,int) as number from test
where type='int') where number 1
will throw an exception
Hi,
application with a H2 db it tells me NO SUITABLE DRIVER
You probably don't have the h2*.jar file in your class path. Could you
check if the jar file is in the class path by calling
Class.forName(org.h2.Driver); in your application?
Regards,
Thomas
Hi,
I have seen that is in priority 2 - can you give some wild guess
about availability of this feature ?
The wild guess is middle of next year (unless somebody else implements
this feature of course).
Regards,
Thomas
--~--~-~--~~~---~--~~
You received this
Hi,
I am sorry to say that, but it looks like a corruption problem. I am
very interested in analyzing and solving this problem. Corruption
problems have top priority for me. I have a few question:
- What is your database URL?
- Did you use multiple connections?
- The first workarounds is:
Hi,
I would have sent you a snippet but we have a lot of other
dependencies which could also probably lead to this unwanted deletion
of tables.
You could append ;TRACE_LEVEL_FILE=3 to the database URL to find out
how data is deleted. Just check the .trace.db file. See also
Hi
I think a trigger is the most database independent solution.
About CHECKSUM TABLE: the MySQL docs say that the fact that two
tables produce the same checksum does not mean that the tables are
identical.
H2 internally keeps a modification counter for each table, I will make
it available in
Hi,
FYI there is now an issue
http://code.google.com/p/h2database/issues/detail?id=46 - progress is
tracked there from now on.
Thanks for your help!
Regards,
Thomas
On Tue, Dec 2, 2008 at 7:21 PM, Milhous [EMAIL PROTECTED] wrote:
Is there a way to tell the create to use VARCHAR(20) instead
Hi,
This is a bug in the Shell tool, it will be fixed in the next release.
A workaround is to use block comments (/* */) instead of line
comments.
I think H2 is more suitable than Derby for Sun's JDK too.
At the time Sun chose Derby, H2 didn't exist yet. Now has IBM reduced
their commitment
Hi,
The connection pool of the Glassfish domain use the mixed mode with
this url : jdbc:h2:./db/slenticketdb;auto_server=true
This means the server listen on a tcp port choosen by itself, right ?
Yes.
When a client like SQuirrel SQL must be connected to the database
server,
it should
Hi again,
The connection pool of the Glassfish domain use the mixed mode with
this url : jdbc:h2:./db/slenticketdb;auto_server=true
Actually there is a problem. You are using a relative path:
./db/...
That's a bit dangerous as it depends on the current working directory,
which is probably
Hi,
When I tried, H2 was about double as fast as HSQLDB. See
http://www.h2database.com/html/performance.html#poleposition_benchmark
- I run the test again today and got this result:
Overall time taken
- HSQLDB: 155'263ms
- H2: 71'275ms
There are a few problems with the PolePosition test:
*
Hi,
Now I'm using LinkedTable objects under following conditions :
ACCESS_MODE=R
OPEN_NEW=TRUE
SHARE_LINKED_CONNECTIONS=TRUE
When I issue SELECT statement which use JOIN clause,
SQLException is thrown because ResultSet is closed under
the condition.
the JDBC I use seems to close all
Hi,
bean id=h2web class=org.h2.tools.Server
factory-method=createWebServer init-method=start destroy-
method=stop
constructor-arg value=-browser /
/bean
createWebServer does not support the argument '-browser'. This is only
supported by the
Hi,
Thanks! I will add a link to your project.
What about calling it a XML to SQL schema converter? Of course
that's still not exact, but quite close.
Regards,
Thomas
On Sun, Dec 7, 2008 at 7:48 PM, trashgod [EMAIL PROTECTED] wrote:
H2 users might be interested in the DBSDL project, which
Hi,
Some more answers:
insert into ... (select ... from)
@rownum:[EMAIL PROTECTED]
I have implemented this now, it will work in the next release.
Query timeouts: statement.setQueryTimeout(15/*seconds*/); is not working
I can't reproduce this problem. What version of H2 do you use? Maybe
Hi,
If you know the jar file name, you can use the zip file mechanism. See also:
http://www.h2database.com/html/features.html#database_in_zip
There is a way to find out the name of the jar file, but it's a bit
tricky. I'm not sure if this works in your case: System.out.println(
Hi,
Thanks a lot for your help! That's great, a patch and a test case! I
will integrate it in the next release.
Regards,
Thomas
On Tue, Dec 9, 2008 at 8:13 PM, Eric Faulhaber [EMAIL PROTECTED] wrote:
Hi Thomas,
The following patch for subversion revision 1121 fixes a scale problem
with
1 - 100 of 4281 matches
Mail list logo