I am currently working on a Servlet which saves songs. I am using hsqldb with JPA and my problem is that on initializing the DB I have to load 10 songs from a JSON file. But as long as I use #GeneratedValue in my song Entity it will just load 5 of the 10 songs.
Here is my code where I access the JSON file.
#Override
public void init() {
emf = Persistence.createEntityManagerFactory("NewPersistenceUnit");
// Problem with loading just half the json
loadSongsFromJSON("songs.json");
System.out.println("DB initialized");
}
private void loadSongsFromJSON(String filename) {
List<Song> songsFromJSON = null;
EntityManager em = null;
try {
ObjectMapper objectMapper = new ObjectMapper();
InputStream is = getClass().getClassLoader().getResourceAsStream(filename);
songsFromJSON = objectMapper.readValue(is, new TypeReference<List<Song>>() {
});
} catch (IOException e) {
e.printStackTrace();
}
if (songsFromJSON != null) {
//for testing
System.out.println("ALL SONGS IN INIT" + System.lineSeparator());
for (Song s : songsFromJSON) {
em = emf.createEntityManager();
em.getTransaction().begin();
//em.persist(s);
em.merge(s);
//for testing
System.out.println("ADDED " + s);
em.getTransaction().commit();
em.close();
}
}
}
All 10 Songs are definitly in the list since:
System.out.println("ADDED " + s);
is printing out all of them.
I also need to use #GeneratedValue so deleting it isn't an option.
Ok I solved the problem.
Tomislav's answer helped me solving it.
The problem was, that in my JSON the songs have a set ID. So I set all ID's to null and then persisted them normally with
s.setId(null);
em.persist(s) //instead of em.merge(s)
Related
i am trying to persist multiple entities to database. but i need to roll back all inserts if one of them faces an exception. how can i do that?
here is what i did:
public class RoleCreationApplyService extends AbstractEntityProxy implements EntityProxy {
#Inject
#Override
public void setEntityManager(EntityManager em) {
super.entityManager = em;
}
#Resource
UserTransaction utx;
public Object acceptAppliedRole(String applyId, Role parentRole, SecurityContext securityContext) throws Exception {
utx.begin();
try {
FilterWrapper filter = FilterWrapper.createWrapperWithFilter("id", Filter.Operator._EQUAL, applyId);
RoleCreationApply roleCreationApply = (RoleCreationApply) getByFilter(RoleCreationApply.class, filter);
Role appliedRole = new Role();
appliedRole.setRoleUniqueName(roleCreationApply.getRoleName());
appliedRole.setRoleName(roleCreationApply.getRoleName());
appliedRole.setRoleDescription(roleCreationApply.getRoleDescription());
appliedRole.setRoleDisplayName(roleCreationApply.getRoleDisplayName());
appliedRole.setCreationTime(new Date());
appliedRole.setCreatedBy(securityContext.getUserPrincipal().getName());
Role childRole = (Role) save(appliedRole);
parentRole.setCreationTime(new Date());
parentRole.setCreatedBy(securityContext.getUserPrincipal().getName());
parentRole = (Role) save(parentRole);
RoleRelation roleRelation = new RoleRelation();
roleRelation.setParentRole(parentRole);
roleRelation.setChildRole(childRole);
RoleRelation savedRoleRelation = (RoleRelation) save(roleRelation);
PostRoleRelation postRoleRelation = new PostRoleRelation();
postRoleRelation.setPost(roleCreationApply.getPost());
postRoleRelation.setRoleRelation(savedRoleRelation);
ir.tamin.framework.domain.Resource result = save(postRoleRelation);
utx.commit();
return result;
} catch (Exception e) {
utx.rollback();
throw new Exception(e.getMessage());
}
}
}
and this is save method in AbstractEntityProxy class:
#Override
#ProxyMethod
public Resource save(Resource clientObject) throws ProxyProcessingException {
checkRelationShips((Entity) clientObject, Method.SAVE, OneToOne.class, ManyToOne.class);
try {
entityManager.persist(clientObject);
} catch (PersistenceException e) {
throw new ResourceAlreadyExistsException(e);
}
return clientObject;
}
but when an exception occures for example Unique Constraint Violated and it goes to catch block, when trying to execute utx.rollback() it complains transaction does not exist and so some entities will persist. but i want all to roll back if one fails.
PS: i don't want to use plain JDBC. what is JPA approach?
I've just begun working with JPA and hibernate, and I have been able to insert query and update on one table, but I cannot seem to delete an entity. Even after I attach the entity with the merge function, the program insists it is detached. Here is the relevant code:
public User getUserByEmail(String email) throws DAOException{
User user = null;
EntityManager em = null; //declare here for use in finally block
try{
//get em for use
EntityManagerFactory factory = JPAUtil.getEntityManagerFactory();
em = factory.createEntityManager();
//setup return type as user
CriteriaBuilder cb = em.getCriteriaBuilder();
CriteriaQuery<User> criteriaQuery = cb.createQuery(User.class);
Root<User> userRoot = criteriaQuery.from(User.class);
//populate the where clause
criteriaQuery.where(cb.equal(userRoot.get("email"), cb.parameter(String.class, "email")));
Query query = em.createQuery(criteriaQuery);
query.setParameter("email", email);
//actually run the query on db
user = (User) query.getSingleResult();
}catch(Exception ex){
DAOException dE = new DAOException(2, "getUserByEmail failed", ex);
//TODO log
throw dE;
}
//cleanup
finally{
if(em != null && em.isOpen()){
em.close();
}
}
return user; //if not found, will still be null
}
#Override
/*
* (non-Javadoc)
* #see dao.IUserDAO#deleteUser(java.lang.String)
*/
public void deleteUser(String email) throws DAOException {
EntityManagerFactory entityManagerFactory;
EntityManager em = null;
EntityTransaction trans = null;
try{
//search user on email
User user = getUserByEmail(email);
//check we found user with specified email
if(user != null){
//setup for interaction with database
entityManagerFactory = JPAUtil.getEntityManagerFactory();
em = entityManagerFactory.createEntityManager();
//alterations of db must occur in scope of transaction
trans = em.getTransaction();
trans.begin();
em.merge(user); //TODO update find to take transaction as parameter
em.remove(user);
trans.commit();
//explicitly flush here
em.flush();
}else{
//TODO we didn't find the user
}
}catch(Exception ex){
DAOException dE = new DAOException(6, "delete failed", ex);
//TODO log
trans.rollback();
throw dE;
}finally{
if(em != null && em.isOpen()){
em.close();
}
}
}
And here is the output from the eclipse console:
"Removing a deattached instance of User#1".
Could someone explain why this bean is still deattached after I explicitly called merge? Also, how would I fix this problem? Thanks.
The content of the specified detached entity object is copied into an
existing managed entity object with the same identity (i.e. same type
and primary key). If the EntityManager does not manage such an entity
object yet a new managed entity object is constructed. The detached
object itself, however, remains unchanged and detached.
Even after merge the detached object itself remains detached, merge will return the new created merged entity
user = em.merge(user); //TODO update find to take transaction as parameter
em.remove(user);
Using Spring Data MongoDB with MongoRepository. I have this bean
#Bean
public Jackson2RepositoryPopulatorFactoryBean repositoryPopulator() {
Jackson2RepositoryPopulatorFactoryBean factory = new Jackson2RepositoryPopulatorFactoryBean();
try {
factory.setResources(resourceResolver.getResources("classpath:static/collections/*.json"));
} catch (IOException e) {
log.error("Could not load data", e);
}
return factory;
}
which just works fine with fongo (db is dropped at the end of a test run) but not with real mongo. If I leave the bean as it is and I switch to real mongo instance, then I get my data base populated but only the first run, if I re-run the project (+tests) then it fails because it's already populated (getting DuplicateKeyException).
How do I populate only on the case the repositories are empty?
Consider using data migration tools like Mongobee. This is basically Liquibase/Flyway for MongoDB.
#Bean
public Jackson2RepositoryPopulatorFactoryBean repositoryPopulator() throws Exception {
Jackson2RepositoryPopulatorFactoryBean factory = new Jackson2RepositoryPopulatorFactoryBean();
try {
Resource[] resources = resourceResolver.getResources("classpath:static/collections/*.json");
//resources to list so I can add only the necessary resources
List<Resource> resourcesToFill = new ArrayList<>();
for (Resource r : resources) {
String collection = r.getFilename().substring(0, r.getFilename().length() - 5);
if (!mongoTemplate().collectionExists(collection))
resourcesToFill.add(r);
}
//back to Array...
resources = new Resource[resourcesToFill.size()];
for(int i=0; i<resources.length; i++)
resources[i] = resourcesToFill.get(i);
factory.setResources(resources); // <-- the reason of this shitty code, why the hell use Array?
} catch (IOException e) {
log.error("Could not load data", e);
}
return factory;
}
I am developing a stand alone application, using sql server compact 3.5 sp2 which runs in process. No Database writes involved. Its purely a reporting application. Read many articles about reusing open db connections in case of sql compact(connection pooling) due to its different behavior from sql server.
Quoting the comments from a quiz opened by Erik Ejlskov Jensen Link, where its discussed an open early close late strategy for sql server compact databases. Based on this, with my limited experience I have implemented a not so complex Connection handling+Data access layer. Basically I am unsure if i am writing it in a recommended way. Please could any one point me in the right direction with rooms for improvement in this connection handling approach i have written?
The DbConnection class
public class FkDbConnection
{
private static SqlCeConnection conn;
private static DataTable table;
private static SqlCeCommand cmd;
~FkDbConnection() { conn = null; }
//This will be called when the main winform loads and connection will be open as long as the main form is open
public static string ConnectToDatabase()
{
try {
conn = new SqlCeConnection(ConfigurationManager.ConnectionStrings["Connstr"].ConnectionString);
if (conn.State == ConnectionState.Closed || conn.State == ConnectionState.Broken)
{
conn.Open();
}
return "Connected";
}
catch(SqlCeException e) { return e.Message; }
}
public static void Disconnect()
{
if (conn.State == ConnectionState.Open || conn.State == ConnectionState.Connecting || conn.State == ConnectionState.Fetching)
{
conn.Close();
conn.Dispose();
//conn = null; //does conn have to be set to null?
}
//else the connection might be already closed due to failure in opening it
else if (conn.State == ConnectionState.Closed) {
conn.Dispose();
//conn = null; //does conn have to be set to null?
}
}
/// <summary>
/// Generic Select DataAccess
/// </summary>
/// <param name="sql"> the sql query which needs to be executed by command object </param>
public static DataTable ExecuteSelectCommand(SqlCeCommand comm)
{
if (conn != null && conn.State == ConnectionState.Open)
{
#region block using datareader
using (table = new DataTable())
{
//using statement needed for reader? Its closed below
using (SqlCeDataReader reader = comm.ExecuteReader())
{
table.Load(reader);
reader.Close(); //is it needed?
}
}
#endregion
# region block using dataadpater
//I read DataReader is faster?
//using (SqlCeDataAdapter sda = new SqlCeDataAdapter(cmd))
//{
// using (table = new DataTable())
// {
// sda.Fill(table);
// }
//}
#endregion
//}
}
return table;
}
/// <summary>
/// Get Data
/// </summary>
/// <param name="selectedMPs"> string csv, generated from a list of selected posts(checkboxes) from the UI, which forms the field names used in SELECT </param>
public static DataTable GetDataPostsCars(string selectedMPs)
{
DataTable dt;
//i know this it not secure sql, but will be a separate question to pass column names to select as parameters
string sql = string.Format(
"SELECT " + selectedMPs + " "+
"FROM GdRateFixedPosts");
using (cmd = new SqlCeCommand(sql,conn))
{
cmd.CommandType = CommandType.Text;
//cmd.Parameters.Add("#fromDateTime",DbType.DateTime);
//cmd.Parameters.Add("#toDateTime",DbType.DateTime);
dt = ExecuteSelectCommand(cmd);
}
return dt;
}
}
The Main UI (Form) in which connection opened, for connection to be open through out. 2 other reporting forms are opened from here. Closing main form closes all, at which point connection is closed and disposed.
private void FrmMain_Load(object sender, EventArgs e)
{
string str = FkDbConnection.ConnectToDatabase();
statStDbConnection.Items[0].Text = str;
}
private void FrmMain_FormClosing(object sender, FormClosingEventArgs e)
{
FkDbConnection.Disconnect();
}
Comments, improvements on this connection class much appreciated. See my questions also inline code
Thank you.
Updated classes as per Erik's suggestion. with a correction on ExecuteSelectCommand() and an additional class which will instantiate command objs in "using" and pass data to the UI. I intent to add separate GetDataForFormX() methods since the dynamic sql for each form may differ. Hope this is ok?
Correction to Erik's code:
public static DataTable ExecuteSelectCommand(SqlCeCommand comm)
{
var table = new DataTable();
if (conn != null && conn.State == ConnectionState.Open)
{
comm.Connection = conn;
using (SqlCeDataReader reader = comm.ExecuteReader())
{
table.Load(reader);
}
}
return table;
}
New FkDataAccess class for passing Data to UI
public class FkDataAccess
{
public static DataTable GetDataPostsCars(string selectedMPs)
{
var table = new DataTable();
string sql = string.Format(
"SELECT " + selectedMPs + " " +
"FROM GdRateFixedPosts");
if (FkDbConnection.conn != null && FkDbConnection.conn.State == ConnectionState.Open)
{
using (SqlCeCommand cmd = new SqlCeCommand(sql, FkDbConnection.conn))
{
cmd.CommandType = CommandType.Text;
//cmd.Parameters.Add("#fromDateTime",DbType.DateTime);
table = FkDbConnection.ExecuteSelectCommand(cmd);
}
}
return table;
}
//public static DataTable GetDataXY(string selectedvals)
// and so on
}
Too much code in your data access class, makes it unreadable and hard to maintain
The SqlCeonnection object will be disposed when you close it (and when the app closes)
You cannot dispose the DataTable if you want to use it elsewhere, and it is an completely managed object anyway.
It is a good pattern to limit your classes to a single responsibility
public class FkDbConnection
{
private static SqlCeConnection conn;
~FkDbConnection() { conn = null; }
//This will be called when the main winform loads and connection will be open as long as the main form is open
public static void ConnectToDatabase()
{
// Handle failure to open in the caller
conn = new SqlCeConnection(ConfigurationManager.ConnectionStrings["Connstr"].ConnectionString);
conn.Open();
}
public static void Disconnect()
{
if (conn != null)
{
conn.Close();
}
}
public static DataTable ExecuteSelectCommand(SqlCeCommand comm)
{
var table = new DataTable();
if (conn != null && conn.State == ConnectionState.Open)
{
comm.Connection = conn;
using (SqlCeDataReader reader = comm.ExecuteReader())
{
table.Load(reader);
}
}
return table;
}
private void FrmMain_Load(object sender, EventArgs e)
{
try
{
FkDbConnection.ConnectToDatabase();
statStDbConnection.Items[0].Text = "Connected";
}
catch (Exception ex)
{
//Inform use that we canot proceed, what she can do to remedy, and exit
}
}
private void FrmMain_FormClosing(object sender, FormClosingEventArgs e)
{
FkDbConnection.Disconnect();
}
I have the following code in UserController in my Session Scoped Bean
public void addItemToBundle(ItemEntity item){
//System.out.println(item.getTitle());
try {
em.getTransaction().begin();
UserEntity user = em.find(UserEntity.class, this.username);
BundleEntity bundle = new BundleEntity();
BundleEntityPK compositePk = new BundleEntityPK();
compositePk.setCheckedOutDate(new Date());
compositePk.setItemId(item.getItemId());
compositePk.setUsername(user.getUsername());
bundle.setId(compositePk);
Set<BundleEntity> bundles = new HashSet<BundleEntity>();
bundles.add(bundle);
user.setBundleEntities(bundles);
em.persist(user);
em.flush();
em.getTransaction().commit();
} finally {
}
}
public String addToBundle(){
try {
addItemToBundle(item);
} catch (NullPointerException e) {
e.getMessage();
}
return null;
}
This code uses private ItemEntity item; which gets passed in by the following JSF markup:
<p:commandLink action="#{itemController.item}">
<f:setPropertyActionListener target="#{itemController.selectedItem}" value="#{movie}" />
</p:commandLink>
(I'm using PrimeFaces in this example) The problem is that the addItemToBundle is not calling any SQL code in the console (I have FINE enabled) and the bundle never gets created or added to the user. I also tried em.persist(user) and em.flush() and setting cascadeType in my UserEntity with no luck.
#OneToMany(mappedBy="userEntity",cascade=CascadeType.PERSIST)
private Set<BundleEntity> bundleEntities;
Thanks!
You know that this:
try {
addItemToBundle(item);
} catch (NullPointerException e) {
e.getMessage();
}
is very bad practice, right? Maybe, that's the problem here, you run into a NPE and never notice it.
You should at least log the exception to know what's going on there (just for demo purposes, I've used stdout, please replace with your favorite logging framework):
try {
addItemToBundle(item);
} catch (NullPointerException e) {
System.err.println(e.getMessage()); //use logger here
}