Danke, Mark.

This IS for a test case with a limited amount of data. So a perfectly valid 
approach, I'd say.

Kay

> On Apr 24, 2015, at 2:14 PM, Mark Struberg <[email protected]> wrote:
> 
> All this is fine for sample apps and smallish installations. 
> 
> However, I personally suggest to neither use the import.sql magic nor the 
> auto-generated schema creation in _real_ production. 
> 
> Why so?
> *) Simple because for a real applciation you will later have to add new 
> features, change the db schema and probably also do a migration. 
> 
> *) For a real application you really like to look at the database. Sometimes 
> you get intermediate mapping tables instead of a simple 1:1 if you don’t take 
> care. My generating the schema upfront you can really look at the create 
> table statements. I alsays use the openjpa-maven-plugin for it. For unit 
> tests I use the sql-maven-plugin to create the test database.
> 
> *) you might like to use liquibase or flywaydb and a schema-version 
> management. I won’t go into details, but check out those 2 projects. 
> Definitly worth looking at.
> 
> LieGrue,
> strub
> 
> 
>> Am 24.04.2015 um 19:11 schrieb Romain Manni-Bucau <[email protected]>:
>> 
>> 2015-04-24 19:01 GMT+02:00 Kay Wrobel <[email protected]>:
>> 
>>> Romain: you're right! Derby.jar doesn't ship with TomEE. No clue how that
>>> ended up in the lib directory. My mistake.
>>> 
>>> So you're suggesting I use Hibernate? But I have a Derby database and I
>>> extracted the data as INSERT statements (from Squirrel tool) and it seems
>>> to contain some Derby-specific stuff, like the notation of date field data
>>> looks like this: {d '2015-04-24'}. Not sure if Hibernate will take that or
>>> complain. I would prefer the other solution where I can provide an SQL
>>> script that is written for that database system (Derby).
>>> 
>>> 
>> import-<name either of the datasource or the persistence unit>.sql is a
>> tomee feature not an hibernate one (import.sql).
>> 
>> PS: if you choose the tomee solution maybe move over tomee@ list
>> 
>> 
>>>> On Apr 24, 2015, at 11:38 AM, Romain Manni-Bucau <[email protected]>
>>> wrote:
>>>> 
>>>> 2015-04-24 18:15 GMT+02:00 Kay Wrobel <[email protected]>:
>>>> 
>>>>> Re your question about the special need: yes. I want to provide a full
>>>>> database, the SAMPLE Derby database that ships with NetBeans/Glassfish
>>> to a
>>>>> user who will be testing a specific use case, but under Apache TomEE.
>>>>> Originally I wanted to provide the entire database as a Derby Database
>>>>> file/directory inside my web app and then connect to it via the Derby
>>>>> Embedded driver. I fail to be able to provide the proper JDBC URL so
>>> that
>>>>> TomEE finds that embedded database in my web app directory (not sure
>>> where
>>>>> TomEE even looks at.
>>>>> 
>>>>> So I decided, well, I would set the database URL to
>>>>> jdbc:derby:memory:sample;create=true as an in-memory database and simply
>>>>> run an SQL script that completely creates the database with CREATE TABLE
>>>>> and INSERT statements. But I didn't know how to instruct the persistence
>>>>> layer to run that script when the application launches. Which is when I
>>>>> found the Oracle reference I mentioned earlier, but it is part of JPA
>>> 2.1
>>>>> as part of the JEE 7 spec. TomEE is based on JEE 6 and JPA 2.0, which
>>>>> apparently doesn't have that feature. Which is why I turned to the
>>> OpenJPA
>>>>> community to see if that is something OpenJPA can do. Apparently this
>>> has
>>>>> been something EclipseLink and Hibernate had figured out a while ago,
>>> but I
>>>>> want to go with what TomEE provides, which is OpenJPA. TomEE also ships
>>>>> with the derby.jar, so there's that.
>>>>> 
>>>>> 
>>>> TomEE doesnt provide derby but hsqldb. In TomEE you can use
>>>> import-<datasource>.sql or import-<persistenceunit>.sql init scripts for
>>>> instance. ALso tomee datasource config (commons-bdcp) has such an option.
>>>> Can be easier alternative to jpa config.
>>>> 
>>>> 
>>>>> Please advise.
>>>>> 
>>>>> Kay
>>>>> 
>>>>>> On Apr 24, 2015, at 10:59 AM, Jody Grassel <[email protected]> wrote:
>>>>>> 
>>>>>> That directive will instruct OpenJPA to introspect the databases to
>>>>> ensure
>>>>>> the tables needed for the table schema defined by your ORM exists, and
>>>>> will
>>>>>> create the table structures automatically if they do not exist.  Do you
>>>>>> have a special need that requires OpenJPA to execute a provided SQL
>>>>> script?
>>>>>> 
>>>>>> On Fri, Apr 24, 2015 at 10:54 AM, Kay Wrobel <[email protected]>
>>> wrote:
>>>>>> 
>>>>>>> Thanks, Jody. How will that let me provide an SQL script containing
>>>>> CREATE
>>>>>>> TABLE and INSERT statements?
>>>>>>> 
>>>>>>> Kay
>>>>>>> 
>>>>>>>> On Apr 24, 2015, at 10:51 AM, Jody Grassel <[email protected]>
>>> wrote:
>>>>>>>> 
>>>>>>>> Add the following property to your persistence unit:
>>>>>>>> 
>>>>>>>> <property name="openjpa.jdbc.SynchronizeMappings"
>>>>>>>> value="buildSchema(ForeignKeys=true)"/>
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> On Fri, Apr 24, 2015 at 10:47 AM, Kay Wrobel <[email protected]>
>>>>> wrote:
>>>>>>>> 
>>>>>>>>> Hi everybody.
>>>>>>>>> 
>>>>>>>>> I am having a rough time finding a way to initialize an Embedded
>>>>>>> In-Memory
>>>>>>>>> Derby database in my web application. I found a reference on
>>> Oracle's
>>>>>>> web
>>>>>>>>> site that states you can initialize a database with DDL and DML
>>>>>>> statements
>>>>>>>>> using properties like the following:
>>>>>>>>> 
>>>>>>>>>>    <properties>
>>>>>>>>>>        <property
>>>>>>>>> name="javax.persistence.schema-generation.database.action"
>>>>>>>>> value="drop-and-create"/>
>>>>>>>>>>        <property
>>>>>>>>> name="javax.persistence.schema-generation.create-source"
>>>>>>> value="script"/>
>>>>>>>>>>        <property
>>>>>>>>> name="javax.persistence.schema-generation.create-script-source"
>>>>>>>>> value="META-INF/sql/create.sql" />
>>>>>>>>>>        <property name="javax.persistence.sql-load-script-source"
>>>>>>>>> value="META-INF/sql/data.sql" />
>>>>>>>>>>    </properties>
>>>>>>>>> 
>>>>>>>>> 
>>>>>>>>> However, that seems to be a new feature in JPA 2.1 spec as part of
>>> JEE
>>>>>>> 6.
>>>>>>>>> I am working with OpenJPA provided by Apache TomEE, which is
>>>>>>> openjpa-2.4.0
>>>>>>>>> non-final, a JPA 2.0 implementation I would imagine. The current
>>>>>>> release on
>>>>>>>>> OpenJPA web site is openjpa 2.3.
>>>>>>>>> 
>>>>>>>>> Is there a way to accomplish this via JPA 2.0  and/or OpenJPA
>>>>>>> properties?
>>>>>>>>> I am trying to initialize an in-memory database for a test case I
>>> try
>>>>> to
>>>>>>>>> provide to someone.
>>>>>>>>> 
>>>>>>>>> Any help would be much appreciated.
>>>>>>>>> 
>>>>>>>>> Kay Wrobel
>>>>>>> 
>>>>>>> 
>>>>> 
>>>>> 
>>> 
>>> 
> 

Reply via email to