RE: HAVING clause
Cool! Just curious, can you elaborate a little as to why you felt that the
Having clause should go in query as opposed to in criteria?
-dave
-Original Message-
From: Jakob Braeuchi [mailto:[EMAIL PROTECTED]
Sent: Monday, March 10, 2003 2:43 PM
To: OJB Users List
Subject: Re: HAVING clause
hi all,
i just commited the first try to support HAVING clause.
this sample code with a having-criteria:
ReportQueryByCriteria query;
Criteria crit, having;
crit = new Criteria();
crit.addGroupBy(new String[] { "id", "name", "vorname" });
having = new Criteria();
having.addGreaterThan("sum(konti.saldo)", new Integer(200));
query = new ReportQueryByCriteria(Person.class, crit);
query.setColumns(new String[] { "id", "name", "vorname",
"sum(konti.saldo)" });
query.setHavingCriteria(having);<<<<
broker.getReportQueryIteratorByQuery(query);;
produces this sql:
SELECT A0.id,A0.name,A0.vorname,sum(A1.saldo) FROM tabPerson A0 INNER
JOIN tabKonto A1 ON A0.id=A1.idPerson GROUP BY A0.id,A0.name,A0.vorname
HAVING sum(A1.saldo) > '200'
do not expect this first version to work perfectly ! there will be
problems when the whereCriteria is empty but the havingCriteria not.
hth
jakob
David Mitchell wrote:
> Is there some particular reason why there's no "addHaving()" method on
>Criteria?
>
> I needed to do the following query on my database. (I am implementing an
>INTERSECTION query, which isnt supported by my database, SQL Server)
>
>SELECT numOperators
>FROM flFactoryOperatorCycleTime
>WHERE (workCenterNdx = 64) AND (productNdx IN (753, 754, 758))
>GROUP BY numOperators
>HAVING (COUNT(*) = 3)
>
> Since I needed to only query a single field (numOperators) and could not
>deal with selecting ALL columns (due to the group by clause), I used a
>report query . So far so good. Then, I ran into trouble because report
>queries can only be created from Criteria, which don't seem to support
>"HAVING" clauses.
>
> My hack (which works) to get around it is to FAKE it! But I do wish I had
a
>better solution.
>
>Criteria crit = new Criteria();
>crit.addEqualTo(IFactoryOperatorCycleTime.WORKCENTERNDX, new
>Integer(((IPersistentObject)workCenter).getNdx()));
>crit.addIn(IFactoryOperatorCycleTime.PRODUCTNDX, productNdxs);
>crit.addSql("(1=1) GROUP BY NUMOPERATORS HAVING (COUNT(*) =
>"+productNdxs.size()+")");
>
> The reason for the "(1=1)" is because when you do an addSql() on a
>criteria, it helpfully puts in the 'AND' part for you, but "AND HAVING..."
>doesnt work,
>and HAVING has to come after the GROUP BY, so if I do the addGroupBy()
>method separately, the HAVING gets put BEFORE the GROUP BY, which also
>doesnt work.
>
> The final query ends up looking like this:
>
>SELECT A0.numOperators FROM flFactoryOperatorCycleTime A0 WHERE ((
>A0.workCenterNdx = ? ) AND
>(A0.productNdx IN ( ? , ? , ? ))) AND (1=1) GROUP BY NUMOPERATORS HAVING
>(COUNT(*) = 3)
>
> Any better suggestions?
>
> thanks-
>David Mitchell
>
>
>
>
>
>
>
>+-+
>This message may contain confidential and/or privileged information. If
you
>are not the addressee or authorized to receive this for the addressee, you
>must not use, copy, disclose or take any action based on this message or
any
>information herein. If you have received this message in error, please
>advise the sender immediately by reply e-mail and delete this message.
>Thank you for your cooperation.
>+-+
>
>-
>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]
+-+
This message may contain confidential and/or privileged information. If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose or take any action based on this message or any
information herein. If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message.
Thank you for your cooperation.
+-+
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
HAVING clause
Is there some particular reason why there's no "addHaving()" method on
Criteria?
I needed to do the following query on my database. (I am implementing an
INTERSECTION query, which isnt supported by my database, SQL Server)
SELECT numOperators
FROM flFactoryOperatorCycleTime
WHERE (workCenterNdx = 64) AND (productNdx IN (753, 754, 758))
GROUP BY numOperators
HAVING (COUNT(*) = 3)
Since I needed to only query a single field (numOperators) and could not
deal with selecting ALL columns (due to the group by clause), I used a
report query . So far so good. Then, I ran into trouble because report
queries can only be created from Criteria, which don't seem to support
"HAVING" clauses.
My hack (which works) to get around it is to FAKE it! But I do wish I had a
better solution.
Criteria crit = new Criteria();
crit.addEqualTo(IFactoryOperatorCycleTime.WORKCENTERNDX, new
Integer(((IPersistentObject)workCenter).getNdx()));
crit.addIn(IFactoryOperatorCycleTime.PRODUCTNDX, productNdxs);
crit.addSql("(1=1) GROUP BY NUMOPERATORS HAVING (COUNT(*) =
"+productNdxs.size()+")");
The reason for the "(1=1)" is because when you do an addSql() on a
criteria, it helpfully puts in the 'AND' part for you, but "AND HAVING..."
doesnt work,
and HAVING has to come after the GROUP BY, so if I do the addGroupBy()
method separately, the HAVING gets put BEFORE the GROUP BY, which also
doesnt work.
The final query ends up looking like this:
SELECT A0.numOperators FROM flFactoryOperatorCycleTime A0 WHERE ((
A0.workCenterNdx = ? ) AND
(A0.productNdx IN ( ? , ? , ? ))) AND (1=1) GROUP BY NUMOPERATORS HAVING
(COUNT(*) = 3)
Any better suggestions?
thanks-
David Mitchell
+-+
This message may contain confidential and/or privileged information. If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose or take any action based on this message or any
information herein. If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message.
Thank you for your cooperation.
+-+
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: Torque XML and Java generation
We're using a customized set of the templates that come with torque-3.0, so I cant tell you what's changed in those since, but I downloaded the "db-torque" cvs project and see that they've split off the "generator" portion (the part we use, which generates schemas and sql). Theres a small mention of it on the torque home page. I had pretty good success using the generator to create ddl code to create a postgres equivalent to our normally-SQLServer application database. I have no idea yet how to copy data from one to the other though (especially since one is at work and one is at home). It did take a little setup time to get all the required folders and files needed by the app though. I copied our xml database schema file to the proper folder, set the build.properties to say "postgresql", and set the postgres profile to point to my database, then ran the "sql" target on the build-torque file that comes from the "conf" subfolder in the torque project. The generated postgres-type sql files, which ended up in "src/sql" were -almost- right for me :) I had to do a couple search-n-replaces on the sql because the code to delete old sequences doesnt use quite the correct name for the sequence. I havent bothered to modify the template, since im not going to be generating this file very often. -dave -Original Message- From: Brian McCallister [mailto:[EMAIL PROTECTED] Sent: Wednesday, February 26, 2003 7:37 AM To: OJB Users List Subject: Torque XML and Java generation I intended to ask this on the Torque list, but as it is bouncing on both db.apache.org and jakarta.apache.org... Does anyone here know the state of the OJB XML and Java class sources from Torque descriptors? The Torque site lists it as "experimental". Thanks, Brian - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: nullable integer fields
YES! Worked like a charm. Thank you very much! -Original Message- From: Thomas Mahler [mailto:[EMAIL PROTECTED] Sent: Friday, February 21, 2003 4:39 PM To: OJB Users List Subject: Re: nullable integer fields Hi David, You'll have to use the Int2IntegerFieldConversion on your Integer attribute. This will convert 0 into null on saving! Details on FieldConversions in jdbc-types.html cheers, Thomas David Mitchell wrote: > I have a table with some foreign key columns defined as integer, nullable. > > Unfortunately, since OJB has mapped my integer columns to "int" values in > Java, I have no way to set a null value on the field! In other words, I > create a new instance of an object, I dont set a value for the null field, > but when I do "save", it blows up because OJB is trying to set the column to > 0, and not null. > > I tried, as an experiment, changing the definitions to Integer from int, > but OJB doesnt appear to like it, and is setting each column to null and > ignoring the columns that actually HAVE a value. im assuming its because of > the INTEGER jdbc-column type set in the repository, perhaps not mapping to > the correct setter and getters for the jdbc code. > > How can I define an OJB integer column to be nullable, and be able to > actually use a null value in that column? > > thanks- > David Mitchell > > > > > +-+ > This message may contain confidential and/or privileged information. If you > are not the addressee or authorized to receive this for the addressee, you > must not use, copy, disclose or take any action based on this message or any > information herein. If you have received this message in error, please > advise the sender immediately by reply e-mail and delete this message. > Thank you for your cooperation. > +-+ > > - > 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] +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
nullable integer fields
I have a table with some foreign key columns defined as integer, nullable. Unfortunately, since OJB has mapped my integer columns to "int" values in Java, I have no way to set a null value on the field! In other words, I create a new instance of an object, I dont set a value for the null field, but when I do "save", it blows up because OJB is trying to set the column to 0, and not null. I tried, as an experiment, changing the definitions to Integer from int, but OJB doesnt appear to like it, and is setting each column to null and ignoring the columns that actually HAVE a value. im assuming its because of the INTEGER jdbc-column type set in the repository, perhaps not mapping to the correct setter and getters for the jdbc code. How can I define an OJB integer column to be nullable, and be able to actually use a null value in that column? thanks- David Mitchell +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
repository questions
I have two questions about the repository definitions: 1- is it possible to tell OJB not to check for the existence of the persistent classes defined in the repository file? My application has several hundred tables, and the full repository file takes quite a while to load on my slow machine every time at application startup. For the moment that only thing I can think of is to remove all table defintions from the file that im not actually using in my code (3/4 of the tables) 2- could collection-descriptors for a set of persistent objects be merged from a second file? My problem is the repository_XXX file is auto-generated, and I havent figured out a good way to make the template insert the collection definitions in the referenced objects automatically. Alternatively, it would be great if I could put those defintions in a separate file from the generated one, then merge the manually-created values after loading the generated values. thanks for any feedback, -dave +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: foreign keys... last time I swear :)
Ok, I guess I pretty much understand it. You use the reference name to do a lookup on the referenced object. So if I want to join two tables on something other than a primary key, I need to use a hand-written sql query. Im okay with that. I believe I will also have to use hand-written sql for my other joins, because I cannot use either lazy-materialization for those foreign objects, nor prefetch that object. I only want to be able to query by those fields without ever actually obtaining the foreign object. For example, if A refers to a B, then in order to make the query "Give me all the As where B's name is X, do I have to really have all those B's instantiated within the A's? I dont want to have to retrieve all those B's from the database unless I actually want the B's. In other words, I cant say "select A.* from A, B where a.b.name='XXX'", unless I have already retrieved the B from the database, to know what it's name is. To answer myself, unless someone corrects my misunderstanding, I have to still write my own SQL query. This is again just fine with me, as long as I know it's the only way to do it. -dave -Original Message- From: Matthew Baird [mailto:[EMAIL PROTECTED]] Sent: Thursday, February 13, 2003 12:26 PM To: OJB Users List Subject: RE: foreign keys... again the docs may not be clear, but the examples are pretty clear. ... +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
report query question
I have a question regarding report queries. What is the purpose of having to provide a OJB persistent class reference when supplying the query code that will be executed? I can see why you need to provide the following String array with the names of the fields you want, but why do I need to pass a specific persistent class type when creating the query? In the java API docs for the ReportQueryByCriteria, the variable is named "targetClass", as if that was the type of object that the query is populating, but it ISNT really, is it? Could I just pass a null in there in lieu of having to pretend that some real OJB object applies to my query? yours, David Mitchell +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: foreign keys... again, joins, and OJB templates
Im having the same issue exactly, and was trying to figure out how to put the question! I have two tables, Product and Group. A product can only be in one group, but a group has many products. Here are representative schemas: table Product ndx int primary key, name varchar, code varchar, groupNdx int table Group ndx int primary key, name varchar, code varchar So, im joining Product.groupNdx with Group.ndx. As I understand it, in the schema definition for Product, you provide a reference descriptor for the link to group, "groupNdx". Im guessing, what happens is that groupNdx is matched with whatever the primary key is on Group, since we've said it refers to the group class in our reference descriptor. But what if you're not joining on a primary key? I see that in the torque schema, it very clearly indicates the local field name and foreign field name in foreign key relationship definitions. Similarly, in the OJB repository, for collections, there is a similar way to provide that foreign field name. But not for reference descriptors..? Is it just an oversight? Finally, what if you're not joining on primary keys at all? What if I want to say "give me all the names of pieces of art, that share an employee's first name?" In torque, it looks like they have addJoin() methods for criteria, so you can link two arbirarily columns in two arbitary tables. How come this doesnt exist for the OJB queries? Oh, one last thing, one request is if the sql, jdbc, ojb-model, and ojb-repository build targets and templates could be brought into the OJB file structure and a "ready to customize for your app ant script" rather than being hardcoded in the torque jarfile... This would allow us (as in the OJB community, rather than the torque community) to take over maintenance of the ojb-related templates if nothing else. yours, David Mitchell -Original Message- From: Robert S. Sfeir [mailto:[EMAIL PROTECTED]] Sent: Thursday, February 13, 2003 10:34 AM To: oJB List Subject: foreign keys... again Still a bit confused, If I have this: ... and I want to setup a foreign key from assignedToID to another table called Users and the column there is called userID I added this as the foreign-key reference: But it doesn't make any sense, if I put assignedToID in the field-ref, how does it know the column I am mapping? If I put the userID like I did in the field-ref how does it know that I am mapping assignedToID? You can't always expect the table column names to match, most of the time they won't. I'm missing something here that would let me tell the descriptiptor. The RequestBO object has a getter and setter which get back a UserBO as suggested earlier, but even if it does a getUserBO() and it gets the UserBO back, UserBO doesn't have a method called getAssignedTOID() nor do I want to do that because that's breaking the OO rules of the framework. I'm totally confused right now and unfortunately, the docs are not very clear. R -- Robert S. Sfeir Senior Java Engineer National Institutes of Health Center for Information Technology Department of Enterprise Custom Applications [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] +-+ This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. +-+ - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: AW: reverse-db (reverse-db2)
try the name of the database user...? or maybe % which is a wildcard for some databases... -Original Message- From: Jeremy [mailto:[EMAIL PROTECTED]] Sent: Friday, February 07, 2003 6:16 PM To: OJB Users List Subject: Re: AW: reverse-db (reverse-db2) Thanks David, now I just need to find the name for a Mysql schema Any clues, anyone? I guess I can find the definitions for this in the jdbc class somewhere. - Again, thanks a bunch for the help...Jeremy David Mitchell wrote: > The schema name it's asking for isnt the schema name that you're generating, > it's a jdbc thing. I believe it's really the name of the user who created > the tables, so different users could have different tables with the same > name. > > Not sure what the setting is for other databases, but for SQL server it > always seems to be "dbo", regardless of who I log in as. "catalog" on the > other hand, seems to map to the name of the database. > > -dave > > > -Original Message- > From: Jeremy [mailto:[EMAIL PROTECTED]] > Sent: Friday, February 07, 2003 2:59 PM > To: OJB Users List > Subject: Re: AW: reverse-db (reverse-db2) > > > Florian, >Thanks for the info, extremely helpful. One more thing that maybe > you could explain a little, now with the CVS version it is asking me for > catalogue regular expression, followed by schema regular expression. I > grok that the catalogue regular expression it is looking for is the > database name, but the schema regular expression escapes me as no schema > is generated yet (that's what this tool does, right? generate schema > from existing data?) so it comes up with an empty schema when I specify > something like a table name or (.*?). Is there anything else I need to > tell it about the schema somewhere? > Thanks, > Jeremy > > Florian Bruckner wrote: > >>Hi Jeremy, >> >>reverse-db2 is not usable yet, this is work in progress and it just does >>reverse engineering of the database, but even this is incomplete and >>generating repository.xml is not working. >> >>Concerning your problem with reverse-db, this is a known issue with MySQL. >>This has been fixed in CVS, so either you can fetch the current source > > from > >>CVS or wait for the next release of OJB, where these changes should be >>included. Reversedb is completely independent of other OJB settings, there >>is no need to set any properties or profiles, you just do a plain connect > > to > >>the database and read the schema, then you can create java stubs and >>repository.xml from it. The reason you are seeing garbage in the reversedb >>tree is simply because it is not initialised on startup. The treemodel > > gets > >>replaced once you hit the "Read Schema" button. This is nothing to worry >>about. >> >>HTH, >> >>Florian >> >> >> >>>-Ursprungliche Nachricht- >>>Von: Jeremy [mailto:[EMAIL PROTECTED]] >>>Gesendet: Freitag, 07. Februar 2003 18:52 >>>An: [EMAIL PROTECTED] >>>Betreff: reverse-db (reverse-db2) >>> >>> >>>Hello, >>> I have been banging my head against how to map out a pre-existing >>>database. I changed the mysql.profile to my mysql db, and changed the >>>build.properties to use the mysql db. Really, this should be >>>superfluous, shouldn't it? I mean, given what documentation on the >>>subject of the reverse-db component, I thought that connecting to a >>>database was all that was necessary (using mysql:jdbc), as it would then >>>read in the database tables and values, then create a schema or map of >>>the database. When I connect to the database, I get no message as to >>>whether it connects or not (using the tests, though, I am able to create >>>the tables for them in the mysql db, and I see traffic, so I believe it >>>is actually connecting), and when I try 'read', I get: >>> >>>[reversedb] java.sql.SQLException: null: Can't read dir of './%/' >>>(Errcode: 2) >>>[reversedb] at com.mysql.jdbc.MysqlIO.nextRow(Unknown Source) >>>[reversedb] at com.mysql.jdbc.MysqlIO.getResultSet(Unknown Source) >>>[reversedb] 12:null: Can't read dir of './%/' (Errcode: 2) >>> >>>Also, why am I seeing in the reverse-db browse window part of the test >>>suite? I mean, before I connect to the db, is it reading in the >>>repository.xml? Do I need to create my own repository.xml or >>>build.xml/build.propertie
RE: AW: reverse-db (reverse-db2)
The schema name it's asking for isnt the schema name that you're generating, it's a jdbc thing. I believe it's really the name of the user who created the tables, so different users could have different tables with the same name. Not sure what the setting is for other databases, but for SQL server it always seems to be "dbo", regardless of who I log in as. "catalog" on the other hand, seems to map to the name of the database. -dave -Original Message- From: Jeremy [mailto:[EMAIL PROTECTED]] Sent: Friday, February 07, 2003 2:59 PM To: OJB Users List Subject: Re: AW: reverse-db (reverse-db2) Florian, Thanks for the info, extremely helpful. One more thing that maybe you could explain a little, now with the CVS version it is asking me for catalogue regular expression, followed by schema regular expression. I grok that the catalogue regular expression it is looking for is the database name, but the schema regular expression escapes me as no schema is generated yet (that's what this tool does, right? generate schema from existing data?) so it comes up with an empty schema when I specify something like a table name or (.*?). Is there anything else I need to tell it about the schema somewhere? Thanks, Jeremy Florian Bruckner wrote: > Hi Jeremy, > > reverse-db2 is not usable yet, this is work in progress and it just does > reverse engineering of the database, but even this is incomplete and > generating repository.xml is not working. > > Concerning your problem with reverse-db, this is a known issue with MySQL. > This has been fixed in CVS, so either you can fetch the current source from > CVS or wait for the next release of OJB, where these changes should be > included. Reversedb is completely independent of other OJB settings, there > is no need to set any properties or profiles, you just do a plain connect to > the database and read the schema, then you can create java stubs and > repository.xml from it. The reason you are seeing garbage in the reversedb > tree is simply because it is not initialised on startup. The treemodel gets > replaced once you hit the "Read Schema" button. This is nothing to worry > about. > > HTH, > > Florian > > >>-Ursprungliche Nachricht- >>Von: Jeremy [mailto:[EMAIL PROTECTED]] >>Gesendet: Freitag, 07. Februar 2003 18:52 >>An: [EMAIL PROTECTED] >>Betreff: reverse-db (reverse-db2) >> >> >>Hello, >> I have been banging my head against how to map out a pre-existing >>database. I changed the mysql.profile to my mysql db, and changed the >>build.properties to use the mysql db. Really, this should be >>superfluous, shouldn't it? I mean, given what documentation on the >>subject of the reverse-db component, I thought that connecting to a >>database was all that was necessary (using mysql:jdbc), as it would then >>read in the database tables and values, then create a schema or map of >>the database. When I connect to the database, I get no message as to >>whether it connects or not (using the tests, though, I am able to create >>the tables for them in the mysql db, and I see traffic, so I believe it >>is actually connecting), and when I try 'read', I get: >> >>[reversedb] java.sql.SQLException: null: Can't read dir of './%/' >>(Errcode: 2) >>[reversedb] at com.mysql.jdbc.MysqlIO.nextRow(Unknown Source) >>[reversedb] at com.mysql.jdbc.MysqlIO.getResultSet(Unknown Source) >>[reversedb] 12:null: Can't read dir of './%/' (Errcode: 2) >> >>Also, why am I seeing in the reverse-db browse window part of the test >>suite? I mean, before I connect to the db, is it reading in the >>repository.xml? Do I need to create my own repository.xml or >>build.xml/build.properties to define a blank db template or something >>for the reverse-db component to fill in? >> >>As an after note, reverse-db2 seems to connect to my database, but using >>the open database like this -->jdbc:mysql://localhost/testdb >>shows me all the databases present, not just testdb, with unknown schema >> (schema not specifies) and new ojb repository gives me that it can't >>find it's OJB.properties file (uses default settings) and it can't find >>it's repository.xml. So what do I put in this repository.xml, I mean, >>it isn't creating a new db, I was hoping it would map a pre-existing one. >>Thanks, >> Jeremy Capps >> >> >>- >>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] +-+ This message may contain confidential and/or privileged information. If you are not th
