Liquibase generates dropdefault change set when nothing has changed - liquibase

I am trying to generate liquibase changeset programtically using liquibase and liquibase-hiberante using the code below
Class.forName("org.mariadb.jdbc.Driver");
Connection connection = DriverManager.getConnection("jdbc:mariadb://localhost:3306/sample_database", "root", "");
connection.getMetaData().getDriverName();
Database database = new MariaDBDatabase();
database.setConnection(new JdbcConnection(connection));
final Liquibase liquibase = new Liquibase("", new ClassLoaderResourceAccessor(), database);
Database hibernateDatabase = new HibernateSpringPackageDatabase();
hibernateDatabase.setDefaultSchemaName("sample_database");
hibernateDatabase.setDefaultCatalogName("sample_database");
hibernateDatabase.setConnection(new JdbcConnection(new HibernateConnection("hibernate:spring:" + "com.example.hibernateddl" + "?dialect=" + MariaDBDialect.class.getName(), new ClassLoaderResourceAccessor())));
ObjectChangeFilter objectChangeFilter = null;
String diffChangeLogFile = "/tmp/abc.xml";
File outFile = new File(diffChangeLogFile);
// liquibase.update(changesToApply, new Contexts(contexts), new LabelExpression(labels));
final Liquibase liquibaseUpdate = new Liquibase(outFile.toString(), new FileSystemResourceAccessor(File.listRoots()), database);
try {
DiffOutputControl diffOutputControl = new DiffOutputControl(false, false, true, null);
diffOutputControl.setObjectChangeFilter(objectChangeFilter);
CommandLineUtils.doDiffToChangeLog(outFile.toString(), hibernateDatabase, database, diffOutputControl,
objectChangeFilter, StringUtil.trimToNull(null),null);
liquibaseUpdate.update("");
} catch (IOException | ParserConfigurationException e) {
throw new LiquibaseException(e);
}
My Entity looks like this
#Entity
public class Sample {
#Id
long id;
#Column(name = "username",columnDefinition = "varchar(15) default 'dfds'")
String username;
#Column(name = "password",columnDefinition = "varchar(15) default null")
String password;
String oneMoreField2;
String oneMoreField3;
}
when program is run first time the changeset is generated correctly and looks like this
<createTable tableName="Sample">
<column name="id" type="BIGINT">
<constraints nullable="false" primaryKey="true" primaryKeyName="SamplePK"/>
</column>
<column name="oneMoreField2" type="VARCHAR(255)"/>
<column name="oneMoreField3" type="VARCHAR(255)"/>
<column name="password" type="VARCHAR(15)"/>
<column defaultValue="dfds" name="username" type="VARCHAR(15)"/>
</createTable>
</changeSet>
when i run the same program second time i expect no changeset to be generated but my changelog file is modified to this
<changeSet author="ArshidDar (generated)" id="1627654919339-1">
<createTable tableName="Sample">
<column name="id" type="BIGINT">
<constraints nullable="false" primaryKey="true" primaryKeyName="SamplePK"/>
</column>
<column name="oneMoreField2" type="VARCHAR(255)"/>
<column name="oneMoreField3" type="VARCHAR(255)"/>
<column name="password" type="VARCHAR(15)"/>
<column defaultValue="dfds" name="username" type="VARCHAR(15)"/>
</createTable>
</changeSet>
<changeSet author="ArshidDar (generated)" id="1627655078245-1">
<dropDefaultValue columnDataType="varchar(255)" columnName="oneMoreField2" tableName="Sample"/>
</changeSet>
<changeSet author="ArshidDar (generated)" id="1627655078245-2">
<dropDefaultValue columnDataType="varchar(255)" columnName="oneMoreField3" tableName="Sample"/>
</changeSet>
<changeSet author="ArshidDar (generated)" id="1627655078245-3">
<addDefaultValue columnDataType="varchar(15)" columnName="password" defaultValue="null" tableName="Sample"/>
</changeSet>
can someone help?

Related

Hibernate: join with 2 NOT primary key

i want to do a left outer join from 2 table, but with 2 key that are not primary key.
This is the native sql query used:
Maybe there is a better way to do it?
SQLQuery query = session.createSQLQuery("
select {A.*},{B.*} from A_TABLE A left outer join B_TABLE B on A.QUOTE_ID = B.QUOTE_NUM ")
.addEntity("A", A_Class.class)
.addJoin("B", "A.bDocs");
for(Object result : query.list())
{
....
}
The mapping A file:
<class name="A_Class" table="A_TABLE" schema="S">
<id name="rowId" type="string">
<column name="ROW_ID" length="60" />
<generator class="assigned" />
</id>
<set name="BDocs" inverse="true" fetch="select" lazy="false">
<key>
<column name="QUOTE_NUM" length="60" not-null="true" />
</key>
<one-to-many class="B_Class" />
</set>
A_Class.java
public class A_Class implements java.io.Serializable
{
private String rowId;
private String QuoteId;
private Set BDocs= new HashSet(0);
// omitted all the set and get
}
The mapping B file:
<hibernate-mapping>
<class name="B" table="B_TABLE" schema="S">
<id name="rowId" type="string">
<column name="ROW_ID" length="60" />
<generator class="assigned" />
</id>
<many-to-one name="A" class="A_Class" fetch="select" lazy="false" outer-join="true" foreign-key="QuoteId" property-ref="QuoteId">
<column name="QUOTE_NUM" length="60" not-null="true" />
</many-to-one>
B_Class.java
public class B_Class implements java.io.Serializable
{
private String rowId;
private String quoteNum;
// omitted all the set and get
}
From this query i obtain 2 objects, one of A type and the other of B Type with a lot of correct datas but the set BDocs in the object of A type isn't filled. The goal is to get only the A Object with the variable BDocs filled with the B Objects.
I don't understand if the problem is in the query or in the mapping files. Anyone can help me?

Cascade ="all" not saving child entities

I think the object model below is saying a Party and PartyName to have a many to one relatioship. An I think the cascade=all i the Party.hbm sshould be having NHib save the child PartyName(s).
But it clearly isn't...
Can someone explain why PartyName isn't being saved with Party and what to do to fix?
Cheers,
Berryl
MAPPING
<class name="Party" table="Parties">
<id name="Id">
<column name="PartyId" />
<generator class="hilo" />
</id>
<discriminator column="Type" not-null="true" type="string" />
<set access="field.camelcase-underscore" cascade="all" inverse="true" name="Names">
<key foreign-key="Party_PartyName_FK">
<column name="PartyNameId" />
</key>
<one-to-many class="Parties.Domain.Names.PartyName, Parties.Domain" />
</set>
<subclass
name="Smack.Core.TestingSupport.NHibernate.TestableDomain.SomeDopeyDomainModel.Student, Smack.Core.TestingSupport"
discriminator-value="Student"
>
<property name="Number" />
<many-to-one
class="Smack.Core.TestingSupport.NHibernate.TestableDomain.SomeDopeyDomainModel.Course, Smack.Core.TestingSupport"
foreign-key="Course_FK"
name="Course">
<column name="CourseId" index="CourseIndex" />
</many-to-one>
</subclass>
<many-to-one access="field.camelcase-underscore" class="Parties.Domain.Party" foreign-key="Party_FK" name="Party">
<column name="PartyId" index="PartyIndex" not-null="true"/>
</many-to-one>
<property name="TheRequiredName" not-null="true" length="50"/>
<property name="EverythingElse" />
<property name="ContextUsed" length="50"/>
<property name="Salutation" length="20"/>
<property name="EffectivePeriod" type="Smack.Core.Data.NHibernate.UserTypes.DateRangeUserType, Smack.Core.Data">
<column name="EffectiveStart"/>
<column name="EffectiveEnd"/>
</property>
Failing Test (and output)
[Test]
public void CanSaveAndLoad_AllProperties()
{
var partyName = NameSeeds.DevName;
partyName.Party = _party;
Assert.That(_party.Names.First(), Is.EqualTo(partyName));
using (var tx = _session.BeginTransaction())
{
_session.Save(_party);
tx.Commit();
}
_session.Clear();
Party foundParty;
using (var tx = _session.BeginTransaction())
{
foundParty = _session.Get<Party>(_party.Id); *** <=== name s/b saved!!
tx.Commit();
}
PartyName foundName = foundParty.Names.First();
//found.Look();
Assert.That(foundName, Is.EqualTo(partyName));
Assert.That(foundName.Party, Is.Not.Null);
Assert.That(foundName.TheRequiredName, Is.EqualTo(partyName.TheRequiredName));
Assert.That(foundName.EverythingElse, Is.EqualTo(partyName.EverythingElse));
Assert.That(foundName.ContextUsed, Is.EqualTo(partyName.ContextUsed));
Assert.That(foundName.Salutation, Is.EqualTo(partyName.Salutation));
Assert.That(foundName.EffectivePeriod, Is.EqualTo(partyName.EffectivePeriod));
}
NHibernate: INSERT INTO Parties (Type, PartyId) VALUES ('Parties.Domain.Party', #p0);#p0 = 32768 [Type: Int32 (0)]
NHibernate: SELECT party0_.PartyId as PartyId2_0_, party0_.Number as Number2_0_, party0_.CourseId as CourseId2_0_, party0_.Type as Type2_0_ FROM Parties party0_ WHERE party0_.PartyId=#p0;#p0 = 32768 [Type: Int32 (0)]
With the mapping of the Names <set> inverse=true, you will have to explicitly call session.Save(partyNameObject) on each member of the collection. If you are looking to have NHibernate automatically save the members of the set when the PartyObject is saved, you need to change the Names <set> inverse attribute to inverse=false. This tells Nhibernate that you want Party to control the relationship between Party and PartyName. You must also remember to add each partyNameObject to the Party.Names collection. Otherwise, they won't be saved when you call Session.Save(partyObject). Keep in mind that having the parent control the relationship may be handy, but if you happen to save the PartyObject without having Loaded the PartyNames collection, NHibernate will update their Party FK to Null. In this scenario with certain Cascade options set on the Names <set>, you might find Nhibernate Deleting them as well.

Nhibernate: On insert new record updating all existing records in table

I'm having a problem with NHibernate. I found this issue in NHibernate Profiler. When I try to insert a new record in the table It's triggering update statement for each record in table.
Like if I have 1000 record in database on insert or update its triggering update statement for each record in database.
Here is my sample
Here is my hbm file and sample code
<property name="ParentCodeId" column="Parent_Code_ID" not-null="false" />
<property name="Code" column="Diagnostic_Code" not-null="false" />
<property name="Descr" column="Description" not-null="true" />
<property name="Level" column="Level" not-null="true" />
<property name="CreatedBy" column="Created_By" not-null="true" />
<property name="CreatedDate" column="Created_Date" not-null="true" />
<property name="ModifiedBy" column="Modified_By" not-null="false" />
<property name="ModifiedDate" column="Modified_Date" not-null="false" />
<property name="IsActive" column="Is_Active" not-null="true" />
var repository = this.DomainObjectRepositoryFactory.Create<DiagnosticCodeRepository, DiagnosticCode>();
repository.BeginTransaction();
diagnosticCode.Id = data.DiagnosticCodeId;
diagnosticCode.ParentCodeId = data.ParentCodeId;
diagnosticCode.Code = data.Code;
diagnosticCode.Descr = data.Descr;
diagnosticCode.IsActive = data.IsActive;
diagnosticCode.Level = data.Level;
// save Diagnostic code
id = repository.Save(diagnosticCode);
repository.Commit();
Any Idea?
Thanks
Imran
Is diagnosticCode a new object?
I don't know if this is your problem, but you might have better luck if you pull the object out of the database, make changes, then save it. You shouldn't ever have to set the id property on your objects by hand.

StoredProcedure returning multiple resultset Sql To Linq using designer

I want to get multiple resultsets from a storedProc using sql to linq. I was not able to generate it from designer so I wrote below code in designer.cs file. But whenever I add something to designer, it refreshes the designer with the markup in .dbml file and hence it removes the below code every time I add something. I have to copy it every time. If I can get corresponding dbml markup for this, it would be great.
[Function(Name = "dbo.GetAllModulesAndOptions")]
[ResultType(typeof(Module))]
[ResultType(typeof(ModuleOption))]
public IMultipleResults GetAllModules()
{
IExecuteResult result = this.ExecuteMethodCall(this, ((MethodInfo)(MethodInfo.GetCurrentMethod())));
return ((IMultipleResults)(result.ReturnValue));
}
I've already defined Module and ModuleOption as tables.
Now when I add below markup in .dbml file it complains
DBML1114: The Name attribute 'Module' of the Type element is already used by another type.
<Function Name="dbo.GetAllModulesAndOptions" Method="GetAllModules">
<ElementType Name="Module">
<Column Name="ModuleId" Type="System.Int64" DbType="BigInt NOT NULL" CanBeNull="false" />
<Column Name="ModuleName" Type="System.String" DbType="VarChar(50)" CanBeNull="true" />
<Column Name="Description" Type="System.String" DbType="VarChar(255)" CanBeNull="true" />
<Column Name="SalesDesc" Type="System.String" DbType="VarChar(MAX)" CanBeNull="true" />
<Column Name="ParentModuleId" Type="System.Int32" DbType="Int" CanBeNull="true" />
</ElementType>
<ElementType Name="ModuleOption">
<Column Name="ModuleOptionId" Type="System.Int32" DbType="Int NOT NULL" CanBeNull="false" />
<Column Name="ModuleOptionName" Type="System.String" DbType="VarChar(50)" CanBeNull="true" />
<Column Name="ModuleOptionDesc" Type="System.String" DbType="VarChar(MAX)" CanBeNull="true" />
<Column Name="DefaultPrice" Type="System.Decimal" DbType="Money" CanBeNull="true" />
<Column Name="ModuleId" Type="System.Int64" DbType="BigInt" CanBeNull="true" />
<Column Name="InUse" Type="System.Int32" DbType="Int" CanBeNull="true" />
</ElementType>
</Function>
I'm using Visual Studio 2008 SP1
Add the method to the partial class for the data context.
You achieve this by adding a file with the same name as the data context alongside the dbml file, and use the class declaration:
public partial class YourDataContext
{
[Function(Name = "dbo.GetAllModulesAndOptions")]
[ResultType(typeof(Module))]
[ResultType(typeof(ModuleOption))]
public IMultipleResults GetAllModules()
{
IExecuteResult result = this.ExecuteMethodCall(this, ((MethodInfo (MethodInfo.GetCurrentMethod())));
return ((IMultipleResults)(result.ReturnValue));
}
}
I'm replying my own answer.
One cannot use the already defined type for result set of stored procedure. So I had to change the name of ElementType to ModuleResult and ModuleOptionResult.
<Function Name="dbo.GetAllModulesAndOptions" Method="GetAllModules">
<ElementType Name="ModuleResult">
<Column Name="ModuleId" Type="System.Int64" DbType="BigInt NOT NULL" CanBeNull="false" />
<Column Name="ModuleName" Type="System.String" DbType="VarChar(50)" CanBeNull="true" />
<Column Name="Description" Type="System.String" DbType="VarChar(255)" CanBeNull="true" />
<Column Name="SalesDesc" Type="System.String" DbType="VarChar(MAX)" CanBeNull="true" />
<Column Name="ParentModuleId" Type="System.Int32" DbType="Int" CanBeNull="true" />
</ElementType>
<ElementType Name="ModuleOptionResult">
<Column Name="ModuleOptionId" Type="System.Int32" DbType="Int NOT NULL" CanBeNull="false" />
<Column Name="ModuleOptionName" Type="System.String" DbType="VarChar(50)" CanBeNull="true" />
<Column Name="ModuleOptionDesc" Type="System.String" DbType="VarChar(MAX)" CanBeNull="true" />
<Column Name="DefaultPrice" Type="System.Decimal" DbType="Money" CanBeNull="true" />
<Column Name="ModuleId" Type="System.Int64" DbType="BigInt" CanBeNull="true" />
<Column Name="InUse" Type="System.Int32" DbType="Int" CanBeNull="true" />
</ElementType>
</Function>
Here are the steps which I took to solve the above problem.
Delete the .designer.cs file
Add above markup to the .dbml file
Exclude .dbml and .dbml.layout file
Include .dbml and .dbml.layout file (this will generate .designer.cs file again but will not include it in the project).
Include .designer file in project.
Get the list of Module type and ModuleOption type as below.
var modules = from row in results.GetResult<ModuleResult>().ToList()
select new Module
{
ModuleId = row.ModuleId,
ModuleName = row.ModuleName,
Description = row.Description,
SalesDesc = row.SalesDesc,
ParentModuleId = row.ParentModuleId
};
var moduleOptions = from row in results.GetResult<ModuleOptionResult>().ToList()
select new ModuleOption
{
ModuleOptionId = row.ModuleOptionId,
ModuleOptionName = row.ModuleOptionName,
ModuleOptionDesc = row.ModuleOptionDesc,
DefaultPrice = row.DefaultPrice,
ModuleId = row.ModuleId,
InUse = row.InUse
};
UPDATE
Still a better way.
Right click the dbml file in solution explorer and select open with. Choose XML Editor and when you save the file within visual studio it automatically generates the designer.cs file.

Querying with NHibernate

I am new to NHibernate and I am trying to learn how to query my data.
Below is the configuration xml. Only the recipe is shown.
I want to be able to query recipes by recipetitle from keywords entered
and also ingredients from ingredientname.
So you might enter "pasta wine" for example.
This is what I have tried but gives me an error.
hql = "from Recipe r " +
"left join r.Images " +
"inner join r.User " +
"inner join r.Ingredients i " +
"where i.IngredientName Like '%pasta%' OR i.IngredientName Like '%wine%' OR r.RecipeTitle Like '%pasta' OR r.RecipeTitle Like '%wine%'";
I want to eager load the collections as well.
Am I going about querying right??
I need to able to build the query string from my search criteria.
This would be easy form me in SQL.
Malcolm
<class name="Recipe" table="Recipes" xmlns="urn:nhibernate-mapping-2.2">
<id name="RecipeID" type="Int32" column="RecipeID">
<generator class="identity" />
</id>
<property name="RecipeTitle" type="String">
<column name="RecipeTitle" />
</property>
<property name="Completed" type="Boolean">
<column name="Completed" />
</property>
<property name="ModifiedOn" type="DateTime">
<column name="ModifiedOn" />
</property>
<property name="Rating" type="Double">
<column name="Rating" />
</property>
<property name="PrepTime" type="Int32">
<column name="PrepTime" />
</property>
<property name="CookTime" type="Int32">
<column name="CookTime" />
</property>
<property name="Method" type="String">
<column name="Method" />
</property>
<bag name="Images" inverse="true" cascade="all">
<key column="RecipeID" />
<one-to-many class="OurRecipes.Domain.RecipeImage, OurRecipes.Domain, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null" />
</bag>
<many-to-one name="Category" column="CategoryID" />
<bag name="Comments" inverse="true" cascade="all">
<key column="RecipeID" />
<one-to-many class="OurRecipes.Domain.Comment, OurRecipes.Domain, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null" />
</bag>
<many-to-one name="User" column="EnteredByID" />
<bag name="Ingredients" inverse="true" cascade="all">
<key column="RecipeID" />
<one-to-many class="OurRecipes.Domain.Ingredient, OurRecipes.Domain, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null" />
</bag>
</class>
To build dynamic queries, I would use the criteria API. This makes the dynamic query much more stable, because you don't need string operations to build it up.
ICriteria query = Session.CreateCriteria(typeof(Recipe), "r")
.CreateCriteria("Ingredients", "i", JoinType.InnerJoin)
.Add(
Expression.Disjunction() // OR
.Add(Expression.Like("i.IngredientName", "%pasta%"))
.Add(Expression.Like("i.IngredientName", "%wine%"))
.Add(Expression.Like("r.RecipeTitle", "%pasta%"))
.Add(Expression.Like("r.RecipeTitle", "%wine%")));
List<Recipe> result = query.List<Recipe>();
Edit:
For eager loading you could set the fetch-mode:
ICriteria query = Session.CreateCriteria(typeof(Recipe), "r")
.SetFetchMode("Images", FetchMode.Join)
.SetFetchMode("Comments", FetchMode.Join)
.SetFetchMode("Ingredients", FetchMode.Join)
But I wouldn't do this because you get the results multiplied by the number of Images, Comments and Ingredients. So if you had 4 Images, 2 Comments and 12 Ingredients, you get your recipe 96 times. You don't recognize this, because NHibernate puts the things together again, but it generates traffic between the application and the database. So better let NHibernate load it with separate queries.
One more edit to show dynamic query composition.
// filter arguments, all are optional and should be omitted if null
List<string> keywords;
TimeSpan? minCookingTime;
TimeSpan? maxCookingTime;
int? minRating;
int? maxRating;
ICriteria query = Session.CreateCriteria(typeof(Recipe), "r");
if (keyword != null)
{
// optional join
query.CreateCriteria("Ingredients", "i", JoinType.InnerJoin);
// add keyword search on ingredientName and RecipeTitle
var disjunction = Expression.Disjunction();
foreach (string keyword in keywords)
{
string pattern = String.Format("%{0}%", keyword);
disjunction
.Add(Expression.Like("i.IngredientName", pattern))
.Add(Expression.Like("r.RecipeTitle", pattern));
}
query.Add(disjunction)
}
if (minCookingTime != null)
{
query.Add(Expression.Ge(r.CookingTime, minCookingTime.Value));
}
if (maxCookingTime != null)
{
query.Add(Expression.Le(r.CookingTime, maxCookingTime.Value));
}
if (minRating != null)
{
query.Add(Expression.Ge(r.Rating, minRating.Value));
}
if (maxRating != null)
{
query.Add(Expression.Le(r.Rating, maxRating.Value));
}
Both Stefan's and Sathish's examples concatenate % operators into the SQL. This is unnecesary as Restrictions.Like (nhib 2.0+) and Expression.Like (before v2.0) have 3 parameter versions with a MatchMode.
Disjunction keywordsCriteria = Restrictions.Disjunction();
foreach (var keyword in keywords)
{
keywordsCriteria.Add(Restrictions.Like("i.IngredientName", keyword, MatchMode.Anywhere));
keywordsCriteria.Add(Restrictions.Like("r.RecipeTitle", keyword, MatchMode.Anywhere));
}
Full text queries are also available with NHibernate Search. See Ayende's example for more details.
Here is the above criteria with dynamic keywords
string searchQuery = "wine pasta";
ICriteria query = Session.CreateCriteria(typeof(Recipe), "r")
.CreateCriteria("Ingredients", "i", JoinType.InnerJoin)
.SetFetchMode("Images", FetchMode.Join)
.SetFetchMode("Comments", FetchMode.Join)
.SetFetchMode("Ingredients", FetchMode.Join)
.SetResultTransformer(new DistinctRootEntityResultTransformer());
var keywords = searchQuery.Split(' ');
Disjunction keywordsCriteria = Restrictions.Disjunction();
foreach (var keyword in keywords)
{
keywordsCriteria.Add(Restrictions.Like("i.IngredientName", string.Format("%{0}%", keyword)));
keywordsCriteria.Add(Restrictions.Like("r.RecipeTitle", string.Format("%{0}%", keyword)));
}
query.Add(keywordsCriteria);
List<Recipe> result = query.List<Recipe>();