hi,I want map mutiple-table to a class.
i.e. TBL_Record_200401, TBL_Record_200402,TBL_Record_200403.......
map to a class , Record.

What should i do?

look at using entity-name in Hibernate3.

With entity-name you can map the same class to multiple table (the thing that now differs is the entity-name instead of classname)

Thanks :)

I need to to split data into diffent tables of a legacy-system.
Depending on the id of an object the mapping should be selecetd.
e.g obj1 with id = 1000 has another mapping than obj2 with id= 2000

<class
...
entity-name ="1000"
...
</class>

<class
...
entity-name ="2000"
</class>

This will result in an error of the JDBC-DRIVER when a SELECT-STATEMENT is issued by hibernate, because the value of entity-name (eg. 1000 or 2000) is used to create an alias for the table name.

Is there any way that the value of the <entity-name> tag is not bounded by the JDBC-Driver?


Hibernate: select 1000x_.id, 1000x_.column1 as column2_0_, 1000x_.column2 as column3_0_ from test.t1 1000x_ where 1000x_.id=?
com.mchange.v2.c3p0.impl.NewPooledConnection@3020ad invalidated by Exception: org.postgresql.util.PSQLException: ERROR: syntax error at or near "x_"

at org.postgresql.util.PSQLException.parseServerError(PSQLException.java:139)

A workaround is to use "n1000" and "n2000" instead of "1000" and "2000". But that´s not sufficient for my problem.

In fact the mapping according to the id of an object is not the whole problem.The reason why I´m trying to omit the JDBC-restrictions is, that I´ve also other properties eg. listId, listSelectionId. Depending on the values of these properties the mapping has to be selected.

e.g.
list1Id = 1 and list1SelectionId =1 => mapping1
list1Id = 1 and list1SelectionId =2 => mapping2
...

And depending on the combination of the values of other properties some values have to be modified.

e.g.
list2Id = 1 and list2SelectionId = 1 => Update .... SET x = y *10
list2Id = 1 and list2SelectionId = 2 => Update .... SET x = y *100
list2Id = 1 and list2SelectionId = 3 => Update .... SET x = y *1000


So entity-name does not all I need.


Any suggestions?

I dont understand why using "n1000" etc doesnt work for you.

all the other stuff you have sounds like its solvable by formula...

Besides your mapping sounds really "scary" ,)

All I found about "formula" is, that it can be used for "select" statements or for the <discriminator> tag when working with the table-per-class-hierarchy mapping strategy.

Maybe I´m wrong, but "formula" can´t be used for update/insert statements.

n1000 works fine for the case of choosing a mapping according to an id.
But when it comes to more complex cases that might not be scalable, because the length of the entity-name i.e. the length of the table alias might be the limiting factor.

yes, sorry - formulas doesnt work with updating (since they per definition is read only)

but the update statement you are showing i really dont see a solution for except with something like batch update.

I cant see how you would map (in any orm) those sqls in a more or less static mapping file. they look more like very specific more or less hand og "non-normalized" algoritmic crafted sql.

/max

[quote="max"]yes, sorry - formulas doesnt work with updating (since they per definition is read only)

but the update statement you are showing i really dont see a solution for except with something like batch update.

I cant see how you would map (in any orm) those sqls in a more or less static mapping file. they look more like very specific more or less hand og "non-normalized" algoritmic crafted sql.

/max[/quote

Yes indeed, "non-normalized" is the right expression for it.
Is there any way to manipulate the update/ insert a runtime (using "PreUpdateEventListener" and "PreInsertEventListener") ?]

No - and if there were, how would you know what sql to execute ?

custompersister is what you need in this case i would assume.//

Thanks a lot for responding so fast.
I will use custompersister and let you know about the results.