Java EE Persistence with Hibernate

Transcription

1 Java EE Persistence with Hibernate Version 2.1 Copyright Capstone Courseware, LLC. All rights reserved.

2 Overview This course works from the very basics of ORM concepts through introductory and intermediate JPA topics. We begin with an introduction to some of what s challenging about ORM, to give a sense of what a JPA provider will be doing for our application code. We then start moving through entity mapping, understanding how to work with the entity manager, and working with JPQL. Then we consider a few more challenging mapping issues, such as inheritance models and composite primary keys, get a look at JPA s support for Java-EE Bean Validation, and conclude by bringing the whole show onto an application server so as to get familiar with practical issues such as data sources and dependency injection, and finally to take advantage of container-managed transactions. Naturally, students with significant SQL experience will have an easier time with this course especially with the JPQL chapter. And those who ve worked with ORMs of almost any nature will have a jump on things as well. But the target is really just Java programmers with enough experience to have some appreciation for the challenges of persistence. JPA 2.x is pretty hot stuff, and most students will find it rewarding to work with JPA almost from the start. Feature by feature, the story keeps building up to a remarkably capable and facile way to do what even just a few years ago was maddening grunt-work. The course goes fairly heavy on lab work, hoping to give students plenty of exercise, especially with querying as there is so much to explore there, and with some of the thornier aspects of persistence operations and mapping strategies. 2

3 Timeline The following breakdowns are approximate, and every class will vary. Day 1 3 hours Chapter 1 2 hours Chapter 2 2½ hours, runs to day 2 Chapter 3 Day 2 2½ hours Chapter 4 3 hours Chapter 5 Day 3 2½ hours Chapter 6 4 hours Chapter 7 Day 4 2½ hours Chapter 8 3½ hours Chapter 9 3

4 Tools Deployed with the Lab Software This course s software requires separate setups of JDK 8, the Crimson text editor, and/or Eclipse. The lab installer sets up the other required tools and libraries, along with the lab software itself. These will all be found in directories under c:/capstone/tools as described in the Tools and Environment page early in the coursebook. Ant Build Process Though most students will either use Eclipse exclusively or will be happy to let the ant command take care of things, some will want to understand the inner workings here a little better. Each web project in the course has its own build.xml and build.properties files; these rely on centralized targets and properties files to define a routine for building the application and for deploying/undeploying as necessary from the server. There are also targets for supporting processes such as creating databases, configuring the server with data sources, message queues, user records, etc. These are all identified as necessary in the coursebook. For a better look under the hood, see that build.xml imports a series of properties and targets files in a central directory Ant. Some of these targets files import each other as well, forming a loose chain. All the logic for building and deploying can be found here. 4

5 Teaching Notes We hope that the coursebook is sufficiently clear and detailed, and so the following notes just capture a few ideas about how to approach a given topic, additional concepts to add to your lectures and discussions, and any surprises you might encounter. Hibernate Logging A general note: we provide a logging.properties file that may be useful in quieting down what we find to be excessive INFO-level-and-higher logging by the Hibernate JPA provider, which can get mighty distracting for console applications such as most of our lab exercises. The Ant-generated run and jpql scripts use these automatically. In Eclipse, on one hand, you may not mind the logging so much since Eclipse color-codes it differently than the console output; but if you want to use it there as well, you will need to set the file as the process logging configuration by way of VM arguments in individual run configurations. Chapter 2 We discuss the alternatives for primary-key generation strategies. The final point, in which we say that IDENTITY can be used with a trigger/sequence combination, is correct as far as the JPA specification is concerned. However, both of our supported providers appear to fall down on this. We use the IDENTITY strategy for all the Earthlings entities, and this is supposed to work cleanly where triggers and sequences are used. But, for Oracle, where we use exactly this approach, both providers try to manipulate the sequence manually and then fail to find it by their best-guess sequence name. We have had to use overrides in orm.xml to effect portability over Derby and Oracle, and have disabled the triggers that would usually be selecting sequence values into the primary-key columns so that the JPA provider can do it directly. You may or may not wish to raise this subject at this point in your class. It may just serve to confuse everyone as we re just getting our feet wet, and perhaps later in the class it ll be worth circling back to it. Or, for more advanced groups, it may make sense to dive right in. Groups using Derby may never need to hear this gory detail. For those using Oracle, it s probably a good idea to mention it no later than your review of Lab 2B, as the overrides appear in orm.xml as of the answer code to this lab. 5

6 Chapter 5 Regarding the footnote on the size function: we have found that neither EclipseLink nor Hibernate generates SQL queries that are quite what you might expect if you were to try to select sizes of more than one collection. The example we discovered is: select size(p.phones), size(p. s) from Person p where p.personid = 1; Of course we were hoping for results such as [3, 2] where person #1 had 3 phone numbers and 2 addresses. But here is the sort of SQL generated by EclipseLink for this query: SELECT COUNT(t0.PHONE_TYPE), COUNT(t2. _TYPE) FROM Pharmacy.PHONES t0, Pharmacy. S t2, Pharmacy.PERSONS t1 WHERE ((t1.person_id = 1) AND ((t0.person_id = t1.person_id) AND (t2.person_id = t1.person_id))) GROUP BY t1.person_id,... This is essentially a three-table join that results in P * E rows for P phone numbers and E addresses. The resulting counts then are both 6 not what we were looking for. And indeed it turns out to be no mean trick to derive this reliably using JPQL. The takeaway seems to be, don t treat size like any other scalar function, and be aware of the pitfalls. Chapter 7 One provider issue is worth noting at this stage. The initial values for most Pharmacy sequences are set by the SQL scripts to be one higher than the maximum ID values currently in the corresponding tables. Which is to say one higher than necessary, since the correct use of these values would be to: read, increment, write, and use as the ID value for an insert. But, we ve found that Hibernate, in its Derby 10.8 dialect, gets this wrong: it reads, grabs the value as the ID for the insert, increments, and writes. Which means that it will trigger a primary-key violation if the sequence is set correctly. So we ve bumped the sequence by one as a workaround. Nevertheless it is possible to encounter this failure we don t recommend doing this! but if you had a Pharmacy application that inserted any type of Person, and you ran it first with EclipseLink and then again with Hibernate... boom. EclipseLink would use the post-increment value of the sequence, and store that. Hibernate would grab that same value and use it, pre-increment, and there s your PK violation. Highly unlikely to come up in class, but something to know. 6

7 Revision History Version 2.1 is the initial release of the course. This is a spinoff of our Course 161, covering in 4 days instead of 5 the same first 7 chapters; and then moving more in the direction of practical issues for Java EE applications, at the expense of some more advanced, pure-jpa topics such as lifecycle, locking, and caching. Troubleshooting If you run into any trouble with code exercises, the first and best thing to do is to doublecheck that the classroom machines have been set up precisely according to the course setup guide. Especially, the wrong version of a tool can cause significant problems; don t wander off-book in this way unless absolutely sure you can support the software that you prefer and that we haven t tested. Check environment variable settings carefully, too; these are the cause of a great many classroom glitches. Below are some specific pitfalls to consider: We provide scripts to run the Derby RDBMS as an application process, and this is usually straightforward, but can be easy to forget especially on day two of the class! The Admin directory is put in the executable path by the SetEnvironment script, so it should be as easy as running StartDerby from wherever you are. It s necessary to know the system password for your Oracle installation. The createdb scripts do this a bit differently one from the other, but both are pretty clear about it and will fail in obvious ways if the password is wrong. To use Oracle, a number of steps are necessary, and these are spelled out in Chapter 1 at the appropriate time for carrying them out. But this can slow class down a bit, because it s easy to miss a step and each depends on the last. One bugaboo in particular for each of Ant and Eclipse usage: o The Ant build generates the run scripts on the fly, writing the full class path into them. This is usually barely noticeable. But if you have already build a given JPA project with Ant, and then run InstallJPAConfigs, the existing run scripts will have class paths that include the wrong JDBC driver JAR. This will show up at runtime as a failure to find the driver class oracle.jdbc.driver.oracledriver. Just run the full build again. o After running InstallJPAConfigs, you may need to refresh your project(s) from Eclipse, so that it picks up on the changes to persistence.xml. 7

8 Errata Here are issues which have been reported since the most recent release: No issues yet! Feedback We very much appreciate whatever feedback we can get on our courseware especially from the instructor s perspective. Naturally, the more specific, the better, and we strongly encourage you to make notes on issues you may encounter in the classroom, whether they re typos, missing files, or suggestions for clearer language to explain a concept. We can t guarantee that we ll act on every suggestion, but we re aggressive about stamping out problems and try to be highly responsive. Hopefully this means that when you give us good feedback, you get a better course the next time you need to teach it. Please direct all courseware feedback to Will Provost Capstone Courseware mailto:provost@capcourse.com For anyone who s interested, we have a very informal defect-tracking system, based in Excel spreadsheets with columns to capture defect location, nature, status, and author feedback. Ultimately, feedback goes into these sheets, so if you want a template, we ll be happy to provide one, to facilitate the reporting process. 8