我在MySQL中有一个名为group_table的表，只有两列user_group_id和group_id（两者都是VARCHAR类型）。这两个列一起形成一个复合主键。
我需要使用子选择IN()执行语句，以根据传递给查询的值列表选择行。
@Override
@SuppressWarnings("unchecked")
public List<GroupTable> getList(List<GroupTable> list)
{
CriteriaBuilder criteriaBuilder=entityManager.getCriteriaBuilder();
CriteriaQuery<GroupTable> criteriaQuery=criteriaBuilder.createQuery(GroupTable.class);
Root<GroupTable> root = criteriaQuery.from(entityManager.getMetamodel().entity(GroupTable.class));
criteriaQuery.where(root.in(list));
return entityManager.createQuery(criteriaQuery).getResultList();
}
该实现产生以下查询。
SELECT group_id,
user_group_id
FROM projectdb.group_table
WHERE ((?, ?) IN ((?, ?), (?, ?)))
/*Binding parameters.*/
bind => [null, null, ROLE_AAA, aaa, ROLE_BBB, aaa]
请注意，有关复合键本身的前两个参数是null。它们应分别为user_group_id和group_id。
为什么不在参数列表中替换它们？
虽然我对在表格中形成复合主键不感兴趣，但这可能是JAAS I用于身份验证的必要条件。
在这种情况下，查询返回的数据库与数据库中提供的列表相同，实际上是不必要的。我实际上需要这个查询来删除多行。
答案 0 :(得分：4)
试试这个
CriteriaBuilder cb = entityManager.getCriteriaBuilder();
CriteriaQuery<GroupTable> cq = cb.createQuery(GroupTable.class);
Root<GroupTable> r = cq.from(GroupTable.class);
Expression<EntityPK> exp = r.get("id"); //EntityPK is your primary composite key class and id is the property name of primary key in GroupTable entity
Predicate predicate = exp.in(list);
cq.select(r).where(predicate);
entityManager.createQuery(cq).getResultList();
我的下表有以下结构
create table EntityA (
col1 integer not null,
col2 integer not null,
description varchar(255),
primary key (col1, col2)
)
以下是实体和复合关键类
@Entity
public class EntityA implements Serializable {
@EmbeddedId
private EntityPK id;
private String description;
// getters, setteres
...........................
............................
}
@Embeddable
public class EntityPK implements Serializable {
private int col1;
private int col2;
// getters, setters, hashcode, equals etc
我的测试代码是
CriteriaBuilder cb = em.getCriteriaBuilder();
CriteriaQuery<EntityA> cq = cb.createQuery(EntityA.class);
Root<EntityA> r = cq.from(EntityA.class);
Expression<EntityPK> exp = r.get("id");
Predicate predicate = exp.in(list);
cq.select(r).where(predicate);
em.createQuery(cq).getResultList();
生成的SQL是
select
entitya0_.col1 as col1_0_,
entitya0_.col2 as col2_0_,
entitya0_.description as descript3_0_
from
EntityA entitya0_
where
entitya0_.col1=?
and entitya0_.col2=?
or entitya0_.col1=?
and entitya0_.col2=?
or entitya0_.col1=?
and entitya0_.col2=?
答案 1 :(得分：3)
这是eclipse链接中缺少的功能。我为这个
开发了一个补丁/** *****************************************************************************
* Copyright (c) 1998, 2013 Oracle and/or its affiliates. All rights reserved.
* This program and the accompanying materials are made available under the
* terms of the Eclipse Public License v1.0 and Eclipse Distribution License v. 1.0
* which accompanies this distribution.
* The Eclipse Public License is available at http://www.eclipse.org/legal/epl-v10.html
* and the Eclipse Distribution License is available at
* http://www.eclipse.org/org/documents/edl-v10.php.
*
* Contributors:
* Oracle - initial API and implementation from Oracle TopLink
* Nicolas Marcotte <nicolas.marcotte@usherbrooke.ca> - patch for IN on composite keys comming from expression builder
***************************************************************************** */
package org.eclipse.persistence.internal.expressions;
import java.io.*;
import java.util.*;
import org.apache.logging.log4j.LogManager;
import org.apache.logging.log4j.Logger;
import org.eclipse.persistence.internal.helper.*;
import org.eclipse.persistence.internal.sessions.AbstractSession;
import org.eclipse.persistence.queries.*;
import org.eclipse.persistence.exceptions.*;
import org.eclipse.persistence.expressions.*;
import org.eclipse.persistence.internal.databaseaccess.*;
import org.eclipse.persistence.internal.sessions.AbstractRecord;
/**
* <p>
* <b>Purpose</b>: Expression SQL printer.
* <p>
* <b>Responsibilities</b>:<ul>
* <li> Print an expression in SQL format.
* <li> Replaces FIELD types with field names from the descriptor.
* <li> Replaces PARAMETER types with row or object values.
* <li> Calls accessor to print primitive types.
* </ul>
* <p>
* @author Dorin Sandu
* @since TOPLink/Java 1.0
*/
public class ExpressionSQLPrinter {
/**
* Stores the current session. The session accessor
* is used to print all the primitive types.
*/
protected AbstractSession session;
/**
* Stores the current platform to access platform specific functions.
*/
protected DatabasePlatform platform;
/**
* Stores the call being created.
*/
protected SQLCall call;
/**
* Stores the row. Used to print PARAMETER nodes.
*/
protected AbstractRecord translationRow;
/**
* Indicates whether fully qualified field names
* (owner + table) should be used or not.
*/
protected boolean shouldPrintQualifiedNames;
// What we write on
protected Writer writer;
/** Used for distincts in functions. */
protected boolean requiresDistinct;
// Used in figuring out when to print a comma in the select line
protected boolean isFirstElementPrinted;
private final ExpressionBuilder builder;
public ExpressionSQLPrinter(AbstractSession session, AbstractRecord translationRow, SQLCall call, boolean printQualifiedNames, ExpressionBuilder builder) {
this.session = session;
this.translationRow = translationRow;
this.call = call;
this.shouldPrintQualifiedNames = printQualifiedNames;
// reference session's platform directly if builder or builder's descriptor is null
if (builder == null || builder.getDescriptor() == null) {
this.platform = getSession().getPlatform();
} else {
this.platform = (DatabasePlatform) getSession().getPlatform(builder.getDescriptor().getJavaClass());
}
this.requiresDistinct = false;
this.builder = builder;
isFirstElementPrinted = false;
}
/**
* Return the call.
*/
public SQLCall getCall() {
return call;
}
/**
* INTERNAL:
* Return the database platform specific information.
*/
public DatabasePlatform getPlatform() {
return this.platform;
}
protected AbstractSession getSession() {
return session;
}
/**
* INTERNAL:
* Return the row for translation
*/
protected AbstractRecord getTranslationRow() {
return translationRow;
}
public Writer getWriter() {
return writer;
}
/**
* INTERNAL:
* Used in figuring out when to print a comma in the select clause
*/
public boolean isFirstElementPrinted() {
return isFirstElementPrinted;
}
public void printExpression(Expression expression) {
translateExpression(expression);
}
public void printField(DatabaseField field) {
if (field == null) {
return;
}
//start of patch 1
//resolve alias if is was not already done
if (builder.getTableAliases() != null) {
DatabaseTable keyAtValue = builder.getTableAliases().keyAtValue(field.getTable());
if (keyAtValue != null) {
field.setTableName(keyAtValue.getName());
}
}
//end of patch 1
try {
// Print the field using either short or long notation i.e. owner + table name.
if (shouldPrintQualifiedNames()) {
getWriter().write(field.getQualifiedNameDelimited(platform));
} else {
getWriter().write(field.getNameDelimited(platform));
}
} catch (IOException exception) {
throw ValidationException.fileError(exception);
}
}
public void printParameter(ParameterExpression expression) {
try {
final Logger logger = LogManager.getLogger();
getCall().appendTranslationParameter(getWriter(), expression, getPlatform(), getTranslationRow());
} catch (IOException exception) {
throw ValidationException.fileError(exception);
}
}
public void printParameter(DatabaseField field) {
getCall().appendTranslation(getWriter(), field);
}
public void printPrimitive(Object value) {
if (value instanceof Collection) {
printValuelist((Collection) value);
return;
}
session.getPlatform().appendLiteralToCall(getCall(), getWriter(), value);
}
public void printNull(ConstantExpression nullValueExpression) {
if (session.getPlatform().shouldBindLiterals()) {
DatabaseField field = null;
Expression localBase = nullValueExpression.getLocalBase();
if (localBase.isFieldExpression()) {
field = ((FieldExpression) localBase).getField();
} else if (localBase.isQueryKeyExpression()) {
field = ((QueryKeyExpression) localBase).getField();
}
session.getPlatform().appendLiteralToCall(getCall(), getWriter(), field);
} else {
session.getPlatform().appendLiteralToCall(getCall(), getWriter(), null);
}
}
public void printString(String value) {
try {
getWriter().write(value);
} catch (IOException exception) {
throw ValidationException.fileError(exception);
}
}
public void printValuelist(Collection values) {
try {
getWriter().write("(");
Iterator valuesEnum = values.iterator();
while (valuesEnum.hasNext()) {
Object value = valuesEnum.next();
// Support nested arrays for IN.
if (value instanceof Collection) {
printValuelist((Collection) value);
} else if (value instanceof Expression) {
((Expression) value).printSQL(this);
//start of patch 2
} else if (value instanceof DatabaseField) {
printExpression(builder.getField((DatabaseField) value));
//end of patch 2
} else {
session.getPlatform().appendLiteralToCall(getCall(), getWriter(), value);
}
if (valuesEnum.hasNext()) {
getWriter().write(", ");
}
}
getWriter().write(")");
} catch (IOException exception) {
throw ValidationException.fileError(exception);
}
}
/*
* Same as printValuelist, but allows for collections containing expressions recursively
*/
public void printList(Collection values) {
try {
getWriter().write("(");
Iterator valuesEnum = values.iterator();
while (valuesEnum.hasNext()) {
Object value = valuesEnum.next();
if (value instanceof Expression) {
((Expression) value).printSQL(this);
} else {
session.getPlatform().appendLiteralToCall(getCall(), getWriter(), value);
}
if (valuesEnum.hasNext()) {
getWriter().write(", ");
}
}
getWriter().write(")");
} catch (IOException exception) {
throw ValidationException.fileError(exception);
}
}
/**
* If a distinct has been set the DISTINCT clause will be printed.
* This is required for batch reading.
*/
public boolean requiresDistinct() {
return requiresDistinct;
}
protected void setCall(SQLCall call) {
this.call = call;
}
/**
* INTERNAL:
* Used in figuring out when to print a comma in the select clause
*/
public void setIsFirstElementPrinted(boolean isFirstElementPrinted) {
this.isFirstElementPrinted = isFirstElementPrinted;
}
/**
* If a distinct has been set the DISTINCT clause will be printed.
* This is required for batch reading.
*/
public void setRequiresDistinct(boolean requiresDistinct) {
this.requiresDistinct = requiresDistinct;
}
protected void setSession(AbstractSession theSession) {
session = theSession;
}
protected void setShouldPrintQualifiedNames(boolean shouldPrintQualifiedNames) {
this.shouldPrintQualifiedNames = shouldPrintQualifiedNames;
}
/**
* INTERNAL:
* Set the row for translation
*/
protected void setTranslationRow(AbstractRecord theRow) {
translationRow = theRow;
}
public void setWriter(Writer writer) {
this.writer = writer;
}
public boolean shouldPrintParameterValues() {
return getTranslationRow() != null;
}
protected boolean shouldPrintQualifiedNames() {
return shouldPrintQualifiedNames;
}
/**
* Translate an expression i.e. call the appropriate
* translation method for the expression based on its
* type. The translation method is then responsible
* for translating the subexpressions.
*/
protected void translateExpression(Expression theExpression) {
theExpression.printSQL(this);
}
}
补丁由 //补丁n的开头和 //补丁n的结尾分隔 我会尝试在上游进行总结，但可能需要一些时间
答案 2 :(得分：1)
您可以使用字段连接方法来解决问题。
创建一个方法，返回要在DTO / Entity中搜索的两个字段。
public String getField1Field2Concatenated() {
return field1+ field2;
}
List<String> ids = list.stream().map(r -> r.getField1Field2Concatenated()).collect(Collectors.toList());
您可以连接两个字段并进行搜索。
Select e from Entity e where concat(e.field1, c.field2) in (:ids)
如果任何字段不是文字，则可以投射
Select e from Entity e where concat(cast(c.field1 as string), c.field2) in (:ids)