| Why is it so SLOW ? [message #171439] |
Mon, 19 June 2006 13:27  |
Eclipse User |
|
|
|
Originally posted by: vincent.tiberghien.ird.sn
Hi,
I have a dataset based on a SQL Query that use a UNION Clause between
two tables and that exit a 237.000 rows table.
In my report, i just add a presentation table with only a footer, i
insert there the expression Total.count(row["my_primary_key"])
Then when i execute the report, it takes years and years, sometimes it
never ends.
The SQL Query executed via MS ACCESS or MYSQL takes about 10 seconds.
I use BIRT 2.0.1
I have tryed connection with a MS ACCESS database through RmiJDBC or the
same database with MySQL and com.mysql.jdbc.Driver (MySQL v3.1), and
it's the same problem.
My computer is a Pentium 4 2GHz, 256MRam
|
|
|
|
|
|
|
|
|
|
| Re: Why is it so SLOW ? [message #172125 is a reply to message #171439] |
Wed, 21 June 2006 07:20  |
| Eclipse User |
|
|
|
Originally posted by: vincent.tiberghien.ird.sn
i thought maybe the SQL Query of my dataset was too long.
so i created a new table in my dataset with the result of my query, then
i just execute a "select *" on that table in my dataset, but my report
is still very very slow to execute.
Has anyone ever succeeded in working with a table of more than 200.000
lines ?
vincent tiberghien a écrit :
> Hi,
> I have a dataset based on a SQL Query that use a UNION Clause between
> two tables and that exit a 237.000 rows table.
> In my report, i just add a presentation table with only a footer, i
> insert there the expression Total.count(row["my_primary_key"])
> Then when i execute the report, it takes years and years, sometimes it
> never ends.
>
> The SQL Query executed via MS ACCESS or MYSQL takes about 10 seconds.
>
> I use BIRT 2.0.1
> I have tryed connection with a MS ACCESS database through RmiJDBC or the
> same database with MySQL and com.mysql.jdbc.Driver (MySQL v3.1), and
> it's the same problem.
> My computer is a Pentium 4 2GHz, 256MRam
|
|
|
|
|
| Re: Why is it so SLOW ? [message #173138 is a reply to message #172072] |
Fri, 23 June 2006 17:57 |
| Eclipse User |
|
|
|
Vincent,
The report design looks fine. Can you please file a bugzilla entry so we can
look into the possible performance problem? Please provide as much as
information as possible to ensure that we reproduce this (e.g., how long is
'years and years' when you run it. If the report never ends do you see any
error messages etc). It's also important to provide your JVM memory setting.
Have you experimented with different VM size settings to see if it changes
things?
Also it's worth pointing out that if at all possible, having your RDBMS do
the aggregation (such as counting). Having BIRT handle the aggregation is
bound to be a lot slower due to the large amount of data rows that must be
passed from the server. If you just want a count of rows create a data set
that does a SELECT COUNT(...) FROM ... and have your data item bound to that
data set.
thanks,
--
Gary Xue
Actuate Corporation - Product Development
BIRT Committer
"vincent tiberghien" <vincent.tiberghien@ird.sn> wrote in message
news:e7b3s6$j6k$1@utils.eclipse.org...
> sorry, here is the correct one
>
> Gary Xue a
|
|
|
Powered by
FUDForum. Page generated in 0.10693 seconds
Home