Cache Question (M:N Relationship)
OJB Gurus, I'm having a problem getting the cache to update in an m:n relationship. I am using the PB API. I have Band entities which represent a band of the musical variety. I also have Event entities which represent a concert/party/banquet at a particular location. Lastly, I have a Performance entity which represents the convergence of a particular Band at a particular Event. It's an m:n relationship between Band and Event, with Performance as the joining table. Here's the problem scenario: View an Event object, which nests a collection of performances. Drill-down to the performance itself, then delete the performance. Prior to the delete, the Event object (containing a nested collection of performances) was cached. I delete the performance via this code (not a complete listing): broker.beginTransaction(); broker.delete(performance); broker.commitTransaction(); After this process succeeds, I check the database. The performance record is gone. Everything goes according to plan EXCEPT, when I return to the Band OR Event screens, the listing for that performance is still there, despite having been removed from the database. As a workaround, I am using a clearCache() call before every query I know exactly how brute force this approach is, but I needed a working fix ASAP Now I'd like to know how to do this (delete an item from a joining table) correctly in OJB. Thank you for any help, Aaron Longwell - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Object Database?
Thomas, Excellent response. I think I'll be spending some time mastering an OODBMS sometime soon. I have a few additional questions of clarification. 1) I have some ambient experience with ODMG and I've read an article on JDO :). Any sites/tutorials/articles you can recommend on introductions to those technologies? I've always used the PB API with OJB to this point. 2) A small percentage of my clients have large enough sites that they require clustered solutions. Do most OODBMS's run in the same JVM as the business logic? Do any provide built in clustering support? DB synchronization amongst databases in a cluster? Thanks once again for your advice, Aaron Thomas Mahler wrote: Hi Aaron, Aaron Longwell wrote: I have only recently realized the awesome benefits of an O/R mapping tool. And I absolutely love OJB! But I've been wondering if OJB is really the best solution for my projects. I typically have 100% control over the environment (small-medium sized projects with no legacy data and no RDBMS requirement). If you are not required to use an RDBMS *and* you are familiar to object persistence APIs like ODMG and JDO, you are typically much better of with an OODBMS! I once was architecting a solution for our company and did the prototyping with a an ODMG compliant ODBMS (Objectivity). I finished a persistent DOM implementation for parsed XML documents within 1 day! Unfortunately the upper management refused to use an OODBMS instead of our existing RDBMS for production. So I thought: It would be great to code against the ODMG API and use the RDBMS as persistence backend. We urgently need an O/R mapping tool with an ODMG API. The rest is history... I'm wondering whether an actual object database might be better for these cases. Can those who have experience give some input on my questions? 1) What Open Source Object databases exist (for Java)? A very cool solution is prevayler.sf.net. It's even cooler than an OODBMS (IMHO). I'm thinking of providing a Prevayler based PersistenceBrokerImpl for some time now... By doing this we could provide all the OJB APIs but without all the RDBMS related hassle. There is also the classic OZONE (also provide an ODMG API) You can search the sourceforge project list for dozens of other implementations 2) How does performance compare to OJB? (I utilize the object cache extensively) prevayler is incredibly fast, as everything is kept in memory and there is no need to perform SQL queries. Everything can be reached by normal Java references. 3) How is development time affected? I'm guessing that systems designed around object databases take a little more front-end planning. My experience with Objectivity and POET is that development time is drastically improved, as you have to think much less about data-management issues. On the other hand you may have some more trouble in the long term because of schema evolution, that is changes to your persistent classes. YOu have to make sure that instances stored before changing the class layout can still be used after applying the changes to the class! This may be really tricky! cheers, Thomas Thanks for your input, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Object Database?
I have only recently realized the awesome benefits of an O/R mapping tool. And I absolutely love OJB! But I've been wondering if OJB is really the best solution for my projects. I typically have 100% control over the environment (small-medium sized projects with no legacy data and no RDBMS requirement). I'm wondering whether an actual object database might be better for these cases. Can those who have experience give some input on my questions? 1) What Open Source Object databases exist (for Java)? 2) How does performance compare to OJB? (I utilize the object cache extensively) 3) How is development time affected? I'm guessing that systems designed around object databases take a little more front-end planning. Thanks for your input, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Urgent: Example for Report Query
I cannot figure out how to do a report query. Here's my set up: Class: Band Fields: Name, ContactPerson, ContactPhone Collection: Performances Class: Performance Fields: Payout, Attendance I need to do a report query that will give me 5 columns: Band.Name, Band.ContactPerson, Band.ContactPhone, Avg(Performance.Payout), Avg(Performance.Attendance). I then need to order by Name, Attendance or Payout based on the user's selection. In real SQL it would look something like this: SELECT Band.Name, AVG(Performance.Payout) As Payout, AVG(Performance.Attendance) As Attendace, Band.ContactPerson, Band.ContactPhone FROM Band INNER JOIN Performance ON Performance.BandID = Band.ID GROUP BY Band.Name, Band.ContactPerson, Band.ContactPhone ORDER BY Attendance ASC I have been thoroughly impressed with the quality of the OJB product, but I have had the hardest time learning how to use it using only the online documentation. I am willing to help out with documentation writing of examples to answer questions like mine above. Thank you for your time, Aaron Longwell - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Urgent: Example for Report Query
Ron,
Thanks for the prompt response... your code looks great... except for a
few items. criteria.addGroupBy() and addOrderBy*() have been
deprecated... I think the same methods on the query object are preferred
now. That's sort of irrelevant though.
I am getting the error: Error reading from result set: Column 'Genre'
not found. Genre is another property on the Band object and a column in
my SQL database. It is irrelevant to this report, but OJB seems to be
hung up trying to populate a Band object with a Genre field that doesn't
exist in the query.
1) Why is OJB creating a Band object this is a report query... isn't
it supposed to return raw data and ignore my Business objects?
2) When this report query finally works, how do I access that data? Is
it a collection of hashmaps keyed by the column name?
Once again, thank you for your prompt and helpfuil response.
Thanks,
Aaron
Ron Gallagher wrote:
Aaron --
This should create the query you need.
String fields[] = new String[]{
/* 0 */ Name,
/* 1 */ avg(Performances.Payout),
/* 2 */ avg(Performances.Attendance),
/* 3 */ ContactPerson,
/* 4 */ ContactPhone};
// Assemble the criteria.
Criteria criteria = new Criteria();
criteria.addGroupBy(fields[0]);
criteria.addGroupBy(fields[3]);
criteria.addGroupBy(fields[4]);
criteria.addOrderByAscending(avg(Performances.Attendance));
// Assemble the query.
ReportQueryByCriteria query =
new ReportQueryByCriteria(
Band.class,
fields,
criteria);
Then, just use the getCollectionByQuery method on the PB to get the results.
HTH
Ron Gallagher
Atlanta, GA
[EMAIL PROTECTED]
From: Aaron Longwell [EMAIL PROTECTED]
Date: 2003/07/21 Mon AM 11:53:46 EDT
To: OJB User [EMAIL PROTECTED]
Subject: Urgent: Example for Report Query
I cannot figure out how to do a report query. Here's my set up:
Class: Band
Fields: Name, ContactPerson, ContactPhone
Collection: Performances
Class: Performance
Fields: Payout, Attendance
I need to do a report query that will give me 5 columns: Band.Name,
Band.ContactPerson, Band.ContactPhone, Avg(Performance.Payout),
Avg(Performance.Attendance).
I then need to order by Name, Attendance or Payout based on the user's
selection.
In real SQL it would look something like this:
SELECT Band.Name, AVG(Performance.Payout) As Payout,
AVG(Performance.Attendance) As Attendace, Band.ContactPerson,
Band.ContactPhone
FROM Band
INNER JOIN Performance ON Performance.BandID = Band.ID
GROUP BY Band.Name, Band.ContactPerson, Band.ContactPhone
ORDER BY Attendance ASC
I have been thoroughly impressed with the quality of the OJB product,
but I have had the hardest time learning how to use it using only the
online documentation. I am willing to help out with documentation
writing of examples to answer questions like mine above.
Thank you for your time,
Aaron Longwell
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Urgent: Example for Report Query
Ron,
Using your logic, I'm trying another (less efficient I know) method of
accomplishing my task. Here's my code:
Criteria criteria = new Criteria();
criteria.addOrderBy(performances.payout, true);
Query query = QueryFactory.newQuery(Band.class, criteria, true);
This code returns Band objects (on which I have an inefficient method to
find average data for the Performance Collections). My code is working
correctly, except OJB is giving me duplicate records. I currently have 5
records in that table, and the first two alphabetically are getting
repeated when I try to order by performances.Payout. This is unusual.
Thanks,
Aaron
Ron Gallagher wrote:
Aaron --
This should create the query you need.
String fields[] = new String[]{
/* 0 */ Name,
/* 1 */ avg(Performances.Payout),
/* 2 */ avg(Performances.Attendance),
/* 3 */ ContactPerson,
/* 4 */ ContactPhone};
// Assemble the criteria.
Criteria criteria = new Criteria();
criteria.addGroupBy(fields[0]);
criteria.addGroupBy(fields[3]);
criteria.addGroupBy(fields[4]);
criteria.addOrderByAscending(avg(Performances.Attendance));
// Assemble the query.
ReportQueryByCriteria query =
new ReportQueryByCriteria(
Band.class,
fields,
criteria);
Then, just use the getCollectionByQuery method on the PB to get the results.
HTH
Ron Gallagher
Atlanta, GA
[EMAIL PROTECTED]
From: Aaron Longwell [EMAIL PROTECTED]
Date: 2003/07/21 Mon AM 11:53:46 EDT
To: OJB User [EMAIL PROTECTED]
Subject: Urgent: Example for Report Query
I cannot figure out how to do a report query. Here's my set up:
Class: Band
Fields: Name, ContactPerson, ContactPhone
Collection: Performances
Class: Performance
Fields: Payout, Attendance
I need to do a report query that will give me 5 columns: Band.Name,
Band.ContactPerson, Band.ContactPhone, Avg(Performance.Payout),
Avg(Performance.Attendance).
I then need to order by Name, Attendance or Payout based on the user's
selection.
In real SQL it would look something like this:
SELECT Band.Name, AVG(Performance.Payout) As Payout,
AVG(Performance.Attendance) As Attendace, Band.ContactPerson,
Band.ContactPhone
FROM Band
INNER JOIN Performance ON Performance.BandID = Band.ID
GROUP BY Band.Name, Band.ContactPerson, Band.ContactPhone
ORDER BY Attendance ASC
I have been thoroughly impressed with the quality of the OJB product,
but I have had the hardest time learning how to use it using only the
online documentation. I am willing to help out with documentation
writing of examples to answer questions like mine above.
Thank you for your time,
Aaron Longwell
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Cascading Delete
I've looked through the docs for this, but they are, forgive me, rather unclear. I have an m:n relationship of Bands and Performances. I want a cascading delete to occur so that whenever the Band is deleted, it's corresponding performances are also deleted. However, when a performance is deleted, I do not want the corresponding Band to be deleted because that Band may have other performances. My Band class-descriptor has a collection descriptor called performances do I put the auto-delete here? Do I need to put an auto-delete anywhere on the performance? I am unclear as to whether the auto-delete attribute says, delete the class-descriptor which contains this attribute when this reference is deleted or if it says, delete this reference if the class-descriptor containing it is deleted. Thanks for the help, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Performance Test Results
Thanks Thomas,
I've stopped running out of connection pools... I'm above the threshold now.
What is the reason for having both a properties and XML configuration
file? I think it would be more convenient to have a globals section in
the repository.xml. Just my opinion.
Thanks again,
Aaron
Thomas Mahler wrote:
Hi Aaron,
OJB uses several Pools. One pool is used for Broker instances.
Then there is another pool for JDBC connections.
What you see is a shortage of JDBC connections.
The Broker pool is configured in OJB.properties, as it is a global
setting.
The Connection Pool is configured in repository.xml, as it can be
configured per JDBC connection.
connection-pool
maxActive=100
whenExhaustedAction=2
/
Allows 100 parallel connnections. If all connections are exhausted the
pool grows (=2).
all features are documented here:
http://db.apache.org/ojb/repository.html#connection-pool
cheers,
Thomas
Aaron Longwell wrote:
I have decided to load test an OJB+Struts web application I am
building. The results, pass! (but not by a whole lot).
My goal was to support 100 simultaneous users. The application is an
event scheduler, for which one view is a calendar-formatted display
of events for the month. It uses OJBs getCollectionByQuery (PB API)
to retrieve the events.
Using Apache's JMeter, I simulated 100 users loading the calendar
page about every 15 seconds. This comes out to around 6-8 requests
per second. Average response times are under 1 second on my very
modest machine. I received no errors after running at this load for
several minutes... in short.. a successful test.
However, upping the users to 150 begins destroying the error
percentage. Every error that occured was a commons-logging unable to
borrow connection from pool exception (see afterword for
StackTrace). Can someone explain the performance enhancement features
of OJB to me?
I have perused the OJB.properties and found maxActive=100 as the
default for the number of brokers. I have changed this to 600 with no
result. In my stack trace I see that there is a
connectionPoolDescriptor setting of maxActive=21 (see below). Is it
possible to change this value in the properties file? What is the
name for the name value pair if so? Are there other enhancements I
can make to support more users?
Thanks,
Aaron Longwell
Selected Portions of Stack Trace:
Used ConnectionManager instance could not obtain a connection: Could
not borrow connection from pool -
org.apache.ojb.broker.metadata.JdbcConnectionDescriptor:
[EMAIL PROTECTED]
jcd-alias=dovedb
default-connection=true
dbms=MySQL
jdbc-level=2.0
driver=com.mysql.jdbc.Driver
protocol=jdbc
sub-protocol=mysql
db-alias=//localhost/dove
eager-release=false
ConnectionPoolDescriptor={whenExhaustedAction=0, maxIdle=-1,
maxActive=21, maxWait=5000, removeAbandoned=false,
numTestsPerEvictionRun=10, testWhileIdle=false,
minEvictableIdleTimeMillis=60, testOnReturn=false,
logAbandoned=false, removeAbandonedTimeout=300,
timeBetweenEvictionRunsMillis=-1, testOnBorrow=true}
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Performance Test Results
Jason,
I am running in Tomcat by itself. I haven't hit any broker instance
errors, just connection pool ones but this is solved after changing
the maxActive setting.
Aaron
Jason McKerr wrote:
You broker instances would still be limited by your connection pool I
think. Are you in a standalone or application server environment?
Jason
On Thu, 2003-07-10 at 22:27, Aaron Longwell wrote:
I have decided to load test an OJB+Struts web application I am building.
The results, pass! (but not by a whole lot).
My goal was to support 100 simultaneous users. The application is an
event scheduler, for which one view is a calendar-formatted display of
events for the month. It uses OJBs getCollectionByQuery (PB API) to
retrieve the events.
Using Apache's JMeter, I simulated 100 users loading the calendar page
about every 15 seconds. This comes out to around 6-8 requests per
second. Average response times are under 1 second on my very modest
machine. I received no errors after running at this load for several
minutes... in short.. a successful test.
However, upping the users to 150 begins destroying the error percentage.
Every error that occured was a commons-logging unable to borrow
connection from pool exception (see afterword for StackTrace). Can
someone explain the performance enhancement features of OJB to me?
I have perused the OJB.properties and found maxActive=100 as the default
for the number of brokers. I have changed this to 600 with no result. In
my stack trace I see that there is a connectionPoolDescriptor setting of
maxActive=21 (see below). Is it possible to change this value in the
properties file? What is the name for the name value pair if so? Are
there other enhancements I can make to support more users?
Thanks,
Aaron Longwell
Selected Portions of Stack Trace:
Used ConnectionManager instance could not obtain a connection: Could not
borrow connection from pool -
org.apache.ojb.broker.metadata.JdbcConnectionDescriptor:
[EMAIL PROTECTED]
jcd-alias=dovedb
default-connection=true
dbms=MySQL
jdbc-level=2.0
driver=com.mysql.jdbc.Driver
protocol=jdbc
sub-protocol=mysql
db-alias=//localhost/dove
eager-release=false
ConnectionPoolDescriptor={whenExhaustedAction=0, maxIdle=-1,
maxActive=21, maxWait=5000, removeAbandoned=false,
numTestsPerEvictionRun=10, testWhileIdle=false,
minEvictableIdleTimeMillis=60, testOnReturn=false,
logAbandoned=false, removeAbandonedTimeout=300,
timeBetweenEvictionRunsMillis=-1, testOnBorrow=true}
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: addLessThan or addBetween with Dates in mySQL
Edison, I'm actually using the same conversion... were you using mySQL? Thanks, Aaron Edson Carlos Ericksson Richter wrote: Hi! I've used Dates normally with less than, greater than and between with SapDB. Unique thing I had todo is put a conversion in field like: conversion=org.apache.ojb.broker.accesslayer.conversions.JavaDate2SqlTimest ampFieldConversion then it works for me. Edson Richter - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB User [EMAIL PROTECTED] Sent: Monday, June 16, 2003 2:42 AM Subject: addLessThan or addBetween with Dates in mySQL OJB Experts, I am using mySQL with OJB to build an event calendar. I need to display 1 month's worth of events. At first, I tried code like this: criteria.addBetween(EventDate, startDate, endDate); where startDate and endDate are java.util.Dates. I got no results, even though there are events in the database between those two dates. I retried with a single addGreaterOrEqualThan(). This worked as expected, returning events whose date was on or after the startDate submitted. I then tried the addLessOrEqualThan(). This returned 0 results, even though there are events before the submitted date in the database. The Javadocs show examples of these commands being used with integers and not dates. Am I wrong to even attempt this? I would prefer not to implement my method an addSQL() call. I would think Date-handling would be automatic with an OJB tool. Thanks for the help, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.489 / Virus Database: 288 - Release Date: 10/6/2003 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: addLessThan or addBetween with Dates in mySQL
Jakob, How do I retrieve the raw SQL for a Query object (or criteria)? The SQL I am using now (working correctly): SimpleDateFormat fmt = (SimpleDateFormat) DateFormat.getDateTimeInstance(); fmt.applyPattern(-MM-dd HH:mm:ss); criteria.addSql(StartDate = ' + fmt.format(startDate) + ' AND StartDate ' + fmt.format(endDate) + '); Query query = QueryFactory.newQuery(Event.class, criteria); Notice that I'm using the addSql() method... I cannot get the addLessThan to work with identical date values. Thanks, Aaron Jakob Braeuchi wrote: hi aaron, afaik ojb does nothing special when using dates with mysql, it maybe the driver. could you please post the sql that caused the problem, so i can check it with my mySQL db. hth jakob Aaron Longwell wrote: Edson, I think there's something wrong with the way OJB is turning an addLessThan to SQL code on the mySQL implementation. I switched my code to SimpleDateFormat the submitted dates and turn them into mySQL-DateTime format (2003-05-18 00:00:00). I then use addSQL(StartDate = ' + strStartDate + ' AND StartDate ' + strEndDate + '); This works just fine... I guess I can deal with it for now. Edson Carlos Ericksson Richter wrote: Nop. I'm using SapDB and MS SQL. Maybe you are having some kind of driver issue here. Edson Richter - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB Users List [EMAIL PROTECTED] Sent: Monday, June 16, 2003 10:40 AM Subject: Re: addLessThan or addBetween with Dates in mySQL Edison, I'm actually using the same conversion... were you using mySQL? Thanks, Aaron Edson Carlos Ericksson Richter wrote: Hi! I've used Dates normally with less than, greater than and between with SapDB. Unique thing I had todo is put a conversion in field like: conversion=org.apache.ojb.broker.accesslayer.conversions.JavaDate2SqlTimes t ampFieldConversion then it works for me. Edson Richter - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB User [EMAIL PROTECTED] Sent: Monday, June 16, 2003 2:42 AM Subject: addLessThan or addBetween with Dates in mySQL OJB Experts, I am using mySQL with OJB to build an event calendar. I need to display 1 month's worth of events. At first, I tried code like this: criteria.addBetween(EventDate, startDate, endDate); where startDate and endDate are java.util.Dates. I got no results, even though there are events in the database between those two dates. I retried with a single addGreaterOrEqualThan(). This worked as expected, returning events whose date was on or after the startDate submitted. I then tried the addLessOrEqualThan(). This returned 0 results, even though there are events before the submitted date in the database. The Javadocs show examples of these commands being used with integers and not dates. Am I wrong to even attempt this? I would prefer not to implement my method an addSQL() call. I would think Date-handling would be automatic with an OJB tool. Thanks for the help, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.489 / Virus Database: 288 - Release Date: 10/6/2003 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.489 / Virus Database: 288 - Release Date: 10/6/2003 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: addLessThan or addBetween with Dates in mySQL
I'm not familiar with p6spy.. is that a class within OJB? Or a log4j-type logger? Jakob Braeuchi wrote: hi aaron, you can use p6spy to log all the sql. btw it works with my installation: QueryByCriteria query; Criteria crit; Date d1, d2; d1 = new Date(70,0,1); d2 = new Date(77,0,1); crit = new Criteria(); crit.addBetween(geburtsDatum, d1, d2); query = new QueryByCriteria(Person.class, crit); query.addOrderByAscending(geburtsDatum); broker.getCollectionByQuery(query); ... field-descriptor name=geburtsDatum column=gebDat jdbc-type=DATE conversion=org.apache.ojb.broker.accesslayer.conversions.JavaDate2SqlDateFieldConversion / ... p6spy output: SELECT A0.isBoss,A0.gebDat,A0.idTest,A0.entryDate,A0.adresse,A0.idPartner,A0.name,A0.test,A0.vorname,A0.id FROM tabPerson A0 WHERE A0.gebDat BETWEEN ? AND ? ORDER BY 2 p6spy output with filed variables: SELECT A0.isBoss,A0.gebDat,A0.idTest,A0.entryDate,A0.adresse,A0.idPartner,A0.name,A0.test,A0.vorname,A0.id FROM tabPerson A0 WHERE A0.gebDat BETWEEN '1970-01-01' AND '1977-01-01' ORDER BY 2 hth jakob Aaron Longwell wrote: Jakob, How do I retrieve the raw SQL for a Query object (or criteria)? The SQL I am using now (working correctly): SimpleDateFormat fmt = (SimpleDateFormat) DateFormat.getDateTimeInstance(); fmt.applyPattern(-MM-dd HH:mm:ss); criteria.addSql(StartDate = ' + fmt.format(startDate) + ' AND StartDate ' + fmt.format(endDate) + '); Query query = QueryFactory.newQuery(Event.class, criteria); Notice that I'm using the addSql() method... I cannot get the addLessThan to work with identical date values. Thanks, Aaron Jakob Braeuchi wrote: hi aaron, afaik ojb does nothing special when using dates with mysql, it maybe the driver. could you please post the sql that caused the problem, so i can check it with my mySQL db. hth jakob Aaron Longwell wrote: Edson, I think there's something wrong with the way OJB is turning an addLessThan to SQL code on the mySQL implementation. I switched my code to SimpleDateFormat the submitted dates and turn them into mySQL-DateTime format (2003-05-18 00:00:00). I then use addSQL(StartDate = ' + strStartDate + ' AND StartDate ' + strEndDate + '); This works just fine... I guess I can deal with it for now. Edson Carlos Ericksson Richter wrote: Nop. I'm using SapDB and MS SQL. Maybe you are having some kind of driver issue here. Edson Richter - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB Users List [EMAIL PROTECTED] Sent: Monday, June 16, 2003 10:40 AM Subject: Re: addLessThan or addBetween with Dates in mySQL Edison, I'm actually using the same conversion... were you using mySQL? Thanks, Aaron Edson Carlos Ericksson Richter wrote: Hi! I've used Dates normally with less than, greater than and between with SapDB. Unique thing I had todo is put a conversion in field like: conversion=org.apache.ojb.broker.accesslayer.conversions.JavaDate2SqlTimes t ampFieldConversion then it works for me. Edson Richter - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB User [EMAIL PROTECTED] Sent: Monday, June 16, 2003 2:42 AM Subject: addLessThan or addBetween with Dates in mySQL OJB Experts, I am using mySQL with OJB to build an event calendar. I need to display 1 month's worth of events. At first, I tried code like this: criteria.addBetween(EventDate, startDate, endDate); where startDate and endDate are java.util.Dates. I got no results, even though there are events in the database between those two dates. I retried with a single addGreaterOrEqualThan(). This worked as expected, returning events whose date was on or after the startDate submitted. I then tried the addLessOrEqualThan(). This returned 0 results, even though there are events before the submitted date in the database. The Javadocs show examples of these commands being used with integers and not dates. Am I wrong to even attempt this? I would prefer not to implement my method an addSQL() call. I would think Date-handling would be automatic with an OJB tool. Thanks for the help, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.489 / Virus Database: 288 - Release Date: 10/6/2003 - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus
Re: Experience in High-Load Environments
Thanks for your advice Mauricio Mauricio CASTRO wrote: Hi Aaron. I don't have any real life example with OJB. I worked in the design and developing of a server written in Java that handled a very heavy load and has a complex data model, it was mounted in a Solaris/sparc machine with Informix. We used object locks, caches and many features that OJB include but homemade. We didn't knew that OJB existed and I doubt OJB in that time was ready for production. The key for success in performance was to improve the design. The cache and precalculated results were very important for performance. On the other hand I am writing a similar software that will use OJB. I have done preliminary tests and I can say that the basics features of OJB behave well for large loads of work. The jdbc driver and the RDBMS have a very important role in performance and stability, and are often underestimated. But the design of the application is paramount for performance and stability. In synthesis I think that OJB can do if you don't try the complex or untested features. I think that in the future, OJB will be very stable an apt for that kind of applications. Mauricio Castro. - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB Users List [EMAIL PROTECTED] Sent: Friday, June 13, 2003 12:20 AM Subject: Re: Experience in High-Load Environments Do you have any examples of projects you've seen/worked on? i.e.: Using Oracle, widgets.com easily handles 500,000 requests an hour I'm actually not that concerned about using OJB for my current project the load isn't that large. But I would like to know the limits of the technology. Mauricio CASTRO wrote: It will depends largely on which RDBMS you use, and which OJB features you use. I think all the basics features of OJB can do the trick. Mauricio Castro. - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB User [EMAIL PROTECTED] Sent: Friday, June 13, 2003 12:14 AM Subject: Experience in High-Load Environments I realize I may be comparing Apples to Enterprise Applications here, but I'd like to hear some feedback about using OJB in high-load (load balanced?) environments. On an OJB web application, how many requests have you seen an application handle? Would anyone be concerned about using OJB on an enormous e-commerce site? (EBay, Amazon, etc)? Thanks for the input, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] . - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Experience in High-Load Environments
Thomas, Thanks for the info. You've answered my questions. I was just curious to hear some info on the impact of OJB's O/R mapping on speed/performance/capacity. I've heard from several people that OJB is efficient enough to use in very high-load situations... and you're very helpful to point out that the performance limits of OJB (however minimal) can be overcome by adding machines to the business tier. Thanks again, Aaron Thomas Mahler wrote: Hi again Aaron, I'm not exactly sure what you expect from the figures that other people provide? Maybe I tell you we once stresstested a Struts/OJB based system with 200 concurrent users firing requests and a sub-second average response time. It was a Tandem Nonstop system with 2 300 Mhz CPU, each equipped with 512MB of Ram. We used the Tandem Nonstop SQL database. This test was helpful for our project, because we got clear figures about our then target environment. But how can such figures be of any help for you? I don't expect that you are running on an environment as exotic as Tandem Nonstop Server :-) By running a Struts/OJB based applications on clustered J2EE Server you will be able to achieve *any* request per minute rate that you want to see. Just add more machines into the cluster. This is possible because J2EE container provide the infrastructure for scalibility and frameworks like STruts and OJB are carefully designed for scalability and not to produce any bottlenecks. So IMO the real question is: how much throughput do you need? And how must your system be sized to give you this throughput. This depends on *many* factors, and OJB is only one of them. So IMO you have to perform such a test on your own! DId you execute our perfoemance testsuite against your target system already? This tool will give you a first impression how OJB APIs compare to native JDBC programming. Aaron Longwell wrote: Jason, Thanks for the feedback. I understand you're probably killing these servers doing this intense processing... but what about # users? Is it one guy running intense simulations... I'm curious about # requests per minute type of loads. Thanks, Aaron Jason McKerr wrote: I've also worked with OJB on high-load situations in J2EE environments. We're using JRun and/or Orion with OJB in a clustered/distributed environment. This is a National Science Foundation project called the Network for Earthquake Engineering Simulation (NEES). The only major problem that we ran into was the cache. JCS just isn't good, and hasn't seemed to get much better over the last year. We ended up plugging in Tangosol's Coherence Clustered Cache into the system. We can also do write-behinds, and buffered data caching that is queued for transaction. That's important to us because we're dealing with very expensive scientific data that _can't_ get lost if a db goes down. Some of these Tsunami experiments can get pretty expensive. Otherwise, we use mostly the PersistenceBroker, and a little of the ODMG. Performance seems better on PB, but less functional. It's not really that much of a problem anyway, because we can cheaply and quickly add app-servers to the cluster. Jason On Thu, 2003-06-12 at 22:14, Aaron Longwell wrote: I realize I may be comparing Apples to Enterprise Applications here, but I'd like to hear some feedback about using OJB in high-load (load balanced?) environments. On an OJB web application, how many requests have you seen an application handle? Would anyone be concerned about using OJB on an enormous e-commerce site? (EBay, Amazon, etc)? Thanks for the input, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Experience in High-Load Environments
Do you have any examples of projects you've seen/worked on? i.e.: Using Oracle, widgets.com easily handles 500,000 requests an hour I'm actually not that concerned about using OJB for my current project the load isn't that large. But I would like to know the limits of the technology. Mauricio CASTRO wrote: It will depends largely on which RDBMS you use, and which OJB features you use. I think all the basics features of OJB can do the trick. Mauricio Castro. - Original Message - From: Aaron Longwell [EMAIL PROTECTED] To: OJB User [EMAIL PROTECTED] Sent: Friday, June 13, 2003 12:14 AM Subject: Experience in High-Load Environments I realize I may be comparing Apples to Enterprise Applications here, but I'd like to hear some feedback about using OJB in high-load (load balanced?) environments. On an OJB web application, how many requests have you seen an application handle? Would anyone be concerned about using OJB on an enormous e-commerce site? (EBay, Amazon, etc)? Thanks for the input, Aaron - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] .
Best Practices: Dates
Goal: Retrieve a collection of objects with dates between X and Y Current Implementation: Event.java (Business Object) contains: java.util.Date startDate repository.xml maps: field-descriptor name=startDate column=StartDate jdbc-type=TIMESTAMP/ PB Query: Criteria criteria = new Criteria(); criteria.addBetween(StartDate, startDate, endDate); Query query = QueryFactory.newQuery(Event.class, criteria); where startDate and endDate are java.util.Dates. When this query executes, I receive no errors about converting from TIMESTAMP to java.util.Date... and I DO receive results. I receive ONLY records where the StartDate column is -00-00 00:00:00... or blank in other words. Why is it matching blank records when I am attempting to find records between date X and date Y? I get similar results when using the addGreaterOrEqualThan submitting startDate. In this case I receive ALL records (including those with -00-00 00:00:00). If you'd rather ignore all the above... I'd like to see best practices for using java.util.Dates with OJB (I'm using mySQL DateTime columns). Thanks, Aaron Longwell - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
SequenceManagerMySQLImpl
I've been looking in the API for an anwer to my previous question about
mySQL auto_increment columns. I found the SequenceManagerMySQLImpl class
and am attempting to implement it.
Here's what I've done:
repository.xml Excerpts:
sequence-manager
className=org.apache.ojb.broker.util.sequence.SequenceManagerMySQLImpl/
class-descriptor class=com.nml.dove.event.EventVO table=event
field-descriptor name=id column=ID
jdbc-type=BIGINT primarykey=true
autoincrement=true/
.
/class-descriptor
In mySQL Database:
ID bigint(20) unsigned not null auto_increment,
PRIMARY KEY ('ID')
When I execute the store command, the record is inserted correctly, but
I get an error in the afterStore() method:
[org.apache.ojb.broker.metadata.fieldaccess.PersistentFieldDefaultImpl]
ERROR: while set field:
object class[ com.nml.dove.event.EventVO
target field: id
target field type: class java.lang.Long
object value class: java.lang.Integer
object value: 3]
[org.apache.ojb.broker.core.PersistenceBrokerImpl]
ERROR: SQLException during SequenceManager.afterStore (for a
com.nml.dove.event.EventVO):
Error setting field:id in object:com.nml.dove.event.EventVO
Error setting field:id in object:com.nml.dove.event.EventVO
I am submitting the EventVO object to OJB (Persistence Broker API) with
all fields except ID populated. I am sending ID as a null value. Because
I am using an auto_increment column in mySQL, I want to store() the
record without an ID.
One clue to this problem is the target field type/object value class
discrepancy above:
target field type: class java.lang.Long
object value class: java.lang.Integer
I have no idea where the Integer is coming from. My mySQL column is a
mediumint, which should correspond to a java Long, right? I have tried
changing the database table to a bigint with no luck.
It appears that the record goes into the database just fine, but that
OJB is trying to update my ValueObject (EventVO) ID column with an
Integer instead of a long.
Is this is bug? Is something wrong with my code?
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: SequenceManagerMySQLImpl
Armin,
Thanks for the quick reply!
Unfortunately, I'm using these mySQL DBs in other apps that aren't able
to utilize OJB.
Have you tried the MySQLImpl before? I am thinking something is wrong
with my setup/code/etc... I am doing something very simple, and I would
think even an experimental Class would work for what I'm trying to do.
Is it difficult to create a sequence manager? I don't need anything
fancy (it doesn't even need to be extents aware).
Thanks,
Aaron
Armin Waibel wrote:
Hi Aaron,
SequenceManagerMySQLImpl is experimental, that's
the reason you could not find it in docs
http://db.apache.org/ojb/sequencemanager.html
In current CVS I add an SequenceManager implementation
called SequenceManagerNativeImpl which supports
native Identity columns. It's also experimental :-(
but seems to work with MySQL.
If possible don't use native identity columns ;-)
regards,
Armin
- Original Message -
From: Aaron Longwell [EMAIL PROTECTED]
To: ObjectRelationalBridge User [EMAIL PROTECTED]
Sent: Thursday, June 05, 2003 5:53 PM
Subject: SequenceManagerMySQLImpl
I've been looking in the API for an anwer to my previous question
about
mySQL auto_increment columns. I found the SequenceManagerMySQLImpl
class
and am attempting to implement it.
Here's what I've done:
repository.xml Excerpts:
sequence-manager
className=org.apache.ojb.broker.util.sequence.SequenceManagerMySQLImpl
/
class-descriptor class=com.nml.dove.event.EventVO table=event
field-descriptor name=id column=ID
jdbc-type=BIGINT primarykey=true
autoincrement=true/
.
/class-descriptor
In mySQL Database:
ID bigint(20) unsigned not null auto_increment,
PRIMARY KEY ('ID')
When I execute the store command, the record is inserted correctly,
but
I get an error in the afterStore() method:
[org.apache.ojb.broker.metadata.fieldaccess.PersistentFieldDefaultImpl]
ERROR: while set field:
object class[ com.nml.dove.event.EventVO
target field: id
target field type: class java.lang.Long
object value class: java.lang.Integer
object value: 3]
[org.apache.ojb.broker.core.PersistenceBrokerImpl]
ERROR: SQLException during SequenceManager.afterStore (for a
com.nml.dove.event.EventVO):
Error setting field:id in object:com.nml.dove.event.EventVO
Error setting field:id in object:com.nml.dove.event.EventVO
I am submitting the EventVO object to OJB (Persistence Broker API)
with
all fields except ID populated. I am sending ID as a null value.
Because
I am using an auto_increment column in mySQL, I want to store() the
record without an ID.
One clue to this problem is the target field type/object value class
discrepancy above:
target field type: class java.lang.Long
object value class: java.lang.Integer
I have no idea where the Integer is coming from. My mySQL column is a
mediumint, which should correspond to a java Long, right? I have tried
changing the database table to a bigint with no luck.
It appears that the record goes into the database just fine, but that
OJB is trying to update my ValueObject (EventVO) ID column with an
Integer instead of a long.
Is this is bug? Is something wrong with my code?
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
mySQL Sequences
I'm working with mySQL 4+ and OJB's current release. I am wondering about the preferred method for autoincrement keys in mySQL. Almost all of my tables have an integer ID field that autoincrements in mySQL. I also edit the databases directly (i.e. not through OJB) through other apps. When I use the default sequence manager, it requires an OJB_HL_SEQ table, and key generation gets a little sloppy when adding or deleting from outside OJB. Here's what I'd like to use: 1) A sequence manager that will use mySQL's built in keying. In other words, I'd like to set the ID field of my (new) object to null, then store the record, and then check to see what ID it got. 2) A sequence manager that does not need additional tables to operate. I sometimes deal with situations where I have only SELECT, UPDATE, DELETE access on the tables I'm utilizing... and I cannot create new tables for OJB to utilize. The documentation seems to suggest that there is a solution to these issues, but I can't find examples to try out. So far though, I love the convenience of OJB. Great work! Thanks, Aaron Longwell - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
