@Convert(converter = StringToListConverter.class) removes the column from generated SQL

[almothafar]

I have nvarchar column that contains strings joined by a delimiter, as for example code1,code2,code3 but in the code, I deal with them as List<String> codes so I used @Convert to deal with conversation, the problem is I see that using this is making the column used to disappear from SQL code generated:

Steps to reproduce

Column

    @Column(name = "TestCodes", length = 300)
    @Convert(converter = StringToListConverter.class)
    private List<String> testCodes;

Converter

@Converter
public class StringToListConverter implements AttributeConverter<List<String>, String> {

    @Override
    public String convertToDatabaseColumn(List<String> list) {
        return list != null ? String.join(",", list) : null;
    }

    @Override
    public List<String> convertToEntityAttribute(String joined) {
        return joined != null ? new ArrayList<>(Arrays.asList(joined.split(","))) : null;
    }
}

[rbygrave]

Note that you can use @DBArray even on databases that don't support ARRAY type. To me that is more logically correct. Ebean doesn't support using @convert in this way and it does provide a lot of limitations in going down that path and we don't have those limitations when we use @DBArray. So to me the expected and better path is to use @DBArray.

…

On Tue, 1 Dec 2020 at 03:18, Al-Mothafar Al-Hasan ***@***.***> wrote: I have nvarchar column that contains strings joined by a delimiter, as for example code1,code2,code3 but in the code, I deal with them as List<String> codes so I used @convert to deal with conversation, the problem is I see that using this is making the column used to disappear from SQL code generated: [image: image] <https://user-images.githubusercontent.com/13887564/100620771-75eadf00-3327-11eb-8e10-7a5a03868983.png> Steps to reproduce Column @column(name = "TestCodes", length = 300) @convert(converter = StringToListConverter.class) private List<String> testCodes; Converter @Converterpublic class StringToListConverter implements AttributeConverter<List<String>, String> { @OverRide public String convertToDatabaseColumn(List<String> list) { return list != null ? String.join(",", list) : null; } @OverRide public List<String> convertToEntityAttribute(String joined) { return joined != null ? new ArrayList<>(Arrays.asList(joined.split(","))) : null; } } — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2119>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABTATJDKSG2OZYQV2RUKS3SSOSR7ANCNFSM4UHUHYZA> .

[rbygrave]

If SQL Server is the only target then another potential option that provides more functionality database wise would be to map as JSON (json array of string). This allows us to use db json functions for predicates like: where testCodes contains 'foo' ... and hence the option of DB functional indexes etc.

…

On Tue, 1 Dec 2020 at 10:17, Rob Bygrave ***@***.***> wrote: Note that you can use @DBArray even on databases that don't support ARRAY type. To me that is more logically correct. Ebean doesn't support using @convert in this way and it does provide a lot of limitations in going down that path and we don't have those limitations when we use @DBArray. So to me the expected and better path is to use @DBArray. On Tue, 1 Dec 2020 at 03:18, Al-Mothafar Al-Hasan < ***@***.***> wrote: > I have nvarchar column that contains strings joined by a delimiter, as > for example code1,code2,code3 but in the code, I deal with them as List<String> > codes so I used @convert to deal with conversation, the problem is I see > that using this is making the column used to disappear from SQL code > generated: > > [image: image] > <https://user-images.githubusercontent.com/13887564/100620771-75eadf00-3327-11eb-8e10-7a5a03868983.png> > Steps to reproduce > > Column > > @column(name = "TestCodes", length = 300) > @convert(converter = StringToListConverter.class) > private List<String> testCodes; > > Converter > > @Converterpublic class StringToListConverter implements AttributeConverter<List<String>, String> { > > @OverRide > public String convertToDatabaseColumn(List<String> list) { > return list != null ? String.join(",", list) : null; > } > > @OverRide > public List<String> convertToEntityAttribute(String joined) { > return joined != null ? new ArrayList<>(Arrays.asList(joined.split(","))) : null; > } > } > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#2119>, or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AABTATJDKSG2OZYQV2RUKS3SSOSR7ANCNFSM4UHUHYZA> > . >

[almothafar]

I tried to use @DbArray and @DbJson the problems are:

• I'm using SQ Server so the array is not an option.
• I "can't"/"don't know how to" set column name, nullable or not, and other attributes except the length.
• IDE reports an error with message
  

[rbygrave]

I'm using SQ Server so the array is not an option.

I don't know why you say that. If we use @DbArray with SQL Server (or any database that doesn't support ARRAY types) Ebean will support it via something that is very similar to the StringToListConverter. So as I see it that is an option.

IDE reports an error with message

You just have to ignore that. JPA has no @DbArray and the IDE warning is going to be JPA orientated and equally unaware.

[almothafar]

Ok the following code seems removing even the error:

    @SuppressWarnings("JpaAttributeTypeInspection")
    @Column(name = "TestCodes")
    @DbArray(length = 300)
    private List<String> testCodes;

But it is varchar, not nvarchar, not a big deal for now:

I'll test it.

[almothafar]

@rbygrave I get error:

[error] - [                    ] a.PostAction - Direct exception java.lang.NoClassDefFoundError: Could not initialize class io.ebean.text.json.EJson
java.util.concurrent.CompletionException: java.lang.NoClassDefFoundError: Could not initialize class io.ebean.text.json.EJson
	at java.base/java.util.concurrent.CompletableFuture.encodeThrowable(CompletableFuture.java:314)
	at java.base/java.util.concurrent.CompletableFuture.completeThrowable(CompletableFuture.java:319)
	at java.base/java.util.concurrent.CompletableFuture$AsyncSupply.run(CompletableFuture.java:1773)
	at play.core.j.HttpExecutionContext.$anonfun$execute$1(HttpExecutionContext.scala:64)
	at akka.dispatch.TaskInvocation.run(AbstractDispatcher.scala:48)
	at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
	at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
	at java.base/java.lang.Thread.run(Thread.java:830)
Caused by: java.lang.NoClassDefFoundError: Could not initialize class io.ebean.text.json.EJson
	at io.ebeaninternal.server.type.ScalarTypeJsonList$Base.bind(ScalarTypeJsonList.java:95)
	at io.ebeaninternal.server.type.ScalarTypeJsonList$Varchar.bind(ScalarTypeJsonList.java:42)
	at io.ebeaninternal.server.type.ScalarTypeJsonList$Base.bind(ScalarTypeJsonList.java:69)
	at io.ebeaninternal.server.deploy.BeanProperty.bind(BeanProperty.java:657)
	at io.ebeaninternal.server.persist.dml.DmlHandler.bindInternal(DmlHandler.java:240)
	at io.ebeaninternal.server.persist.dml.DmlHandler.bind(DmlHandler.java:212)
	at io.ebeaninternal.server.persist.dmlbind.BindableProperty.dmlBind(BindableProperty.java:54)
	at io.ebeaninternal.server.persist.dmlbind.BindableList.dmlBind(BindableList.java:66)
	at io.ebeaninternal.server.persist.dml.InsertMeta.bind(InsertMeta.java:149)
	at io.ebeaninternal.server.persist.dml.InsertHandler.bind(InsertHandler.java:98)

[rbygrave]

Caused by: java.lang.NoClassDefFoundError: Could not initialize class io.ebean.text.json.EJson

So this is suggesting that either EJson is not in the class path or that something is interesting with class loading (what is special/different about the class loading mechanism).

The first question is - Is io.ebean.text.json.EJson in the class path? (It is in the ebean-api module)

[almothafar]

It is there:

But in runtime I tried also something like:

        try {
            Class<?> clazz = Class.forName("io.ebean.text.json.EJson");
        } catch (NoClassDefFoundError e) {
            logger.error("io.ebean.text.json.EJson not found");
        }

And logger did log it:

[error] - [                    ] m.StartupModule - io.ebean.text.json.EJson not found

Note that if I set testCodes with null, the API works fine and the data got stored.

[almothafar]

@rbygrave
So it is not OK even to use @Converter with ebean? because I'm getting wired issue if I tried to use it:

java.util.concurrent.CompletionException: java.lang.NullPointerException
	at java.base/java.util.concurrent.CompletableFuture.encodeThrowable(CompletableFuture.java:314)
	at java.base/java.util.concurrent.CompletableFuture.completeThrowable(CompletableFuture.java:319)
	at java.base/java.util.concurrent.CompletableFuture$UniApply.tryFire(CompletableFuture.java:645)
	at java.base/java.util.concurrent.CompletableFuture$Completion.run(CompletableFuture.java:478)
	at play.core.j.HttpExecutionContext.$anonfun$execute$1(HttpExecutionContext.scala:64)
	at akka.dispatch.TaskInvocation.run(AbstractDispatcher.scala:48)
	at akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinTask.exec(ForkJoinExecutorConfigurator.scala:48)
	at java.base/java.util.concurrent.ForkJoinTask.doExec(ForkJoinTask.java:290)
	at java.base/java.util.concurrent.ForkJoinPool$WorkQueue.topLevelExec(ForkJoinPool.java:1016)
	at java.base/java.util.concurrent.ForkJoinPool.scan(ForkJoinPool.java:1665)
Caused by: java.lang.NullPointerException: null
	at io.ebeaninternal.server.deploy.BeanPropertyAssocMany.createReference(BeanPropertyAssocMany.java:636)
	at io.ebeaninternal.server.deploy.BeanDescriptor.lazyLoadMany(BeanDescriptor.java:2319)
	at io.ebeaninternal.server.deploy.BeanDescriptor.lazyLoadMany(BeanDescriptor.java:2302)
	at io.ebeaninternal.server.loadcontext.DLoadBeanContext$LoadBuffer.loadBean(DLoadBeanContext.java:197)
	at io.ebean.bean.EntityBeanIntercept.loadBeanInternal(EntityBeanIntercept.java:852)
	at io.ebean.bean.EntityBeanIntercept.loadBean(EntityBeanIntercept.java:829)
	at io.ebean.bean.EntityBeanIntercept.preGetter(EntityBeanIntercept.java:923)
	at m.s.XEntity._ebean_get_testCodes(XEntity.java:6)
	at m.s.XEntity.getTestCodes(XEntity.java:356)
	at v2.s.Resource.<init>(Resource.java:59)

where in Resource.java:59 I just access to data like data.getTestCodes()

So now Can't use @DbArray and @Converter :)

my current workaround is doing this:

    @Column(name = "TestCodes", length = 300)
    private String testCodes;

    public List<String> getTestCodes() {
        return testCodes != null ? List.of(testCodes.split(",")) : null;
    }

    public LocationSchedulePreferenceData setTestCodes(List<String> testCodes) {
        this.testCodes = testCodes != null ? String.join(",", testCodes) : null;
        return this;
    }

[rbygrave]

So it is not OK even to use @converter with ebean? because I'm getting wired issue if I tried to use it:

Have we seen the code / mapping that produce that error?

So now Can't use @DBArray and @converter :)

I don't understand this comment - it isn't clear to me what this means.

my current workaround is doing this:

Sure, but @DbArray would also work too so I think we are missing some information. Ideally we see a failing test case without any dependency on Akka or anything (and then compare it to the existing tests to see if there is missing coverage or something).

[rbygrave]

Forking and submitting a PR on https://github.com/ebean-orm-examples/example-minimal is a good way to provide a failing test (if we think we have a bug).

[almothafar]

@rbygrave Yeah it saves a lot of time to provide a working project having an issue, already got a minimal project https://github.com/almothafar/play-java-ebean-sql-server but I barely have time, I'm under infinite stress these days (well actually years....)...
I'll try to maintain that project to reproduce this as soon as I can breathe...

[rbygrave]

I think we will close this for now. Re-open if there is more but I think we should be just using @DBArray or @dbjson