ServuceImpl的saveOrUpdateBatch中校验是否需要更新是单条查询造成速度慢的优化

[sgps000]

请详细描述需要增加的功能
版本：3.5.4.1
环境：JDK17

想要类似以下的优化实现，将单条校验改为批量查询校验：

public boolean saveOrUpdateBatch(Collection<T> entityList, int batchSize) {
        if (CollectionUtils.isEmpty(entityList)) return false;
        TableInfo tableInfo = TableInfoHelper.getTableInfo(entityClass);
        Assert.notNull(tableInfo, "error: can not execute. because can not find cache of TableInfo for entity!");
        String keyProperty = tableInfo.getKeyProperty();
        Assert.notEmpty(keyProperty, "error: can not execute. because can not find column for id from entity!");

        String sqlStatement = SqlHelper.getSqlStatement(this.mapperClass, SqlMethod.INSERT_ONE);
        Consumer<SqlSession> consumer = sqlSession -> {
            int size = entityList.size();
            int idxLimit = Math.min(batchSize, size);
            int i = 1;
            // 批处理临时集合
            List<T> batchList = new ArrayList<>();

            for (T element : entityList) {
                batchList.add(element);

                // 批处理
                if (i == idxLimit) {
                    // 批数据id集合
                    List<Object> idList = batchList.stream().map(e -> tableInfo.getPropertyValue(e, keyProperty))
                            .filter(StringUtils::checkValNotNull).toList();

                    // 查询批数据中是更新的id集合
                    List<Object> updateIdList = new ArrayList<>();
                    if (!idList.isEmpty()) {
                        // todo 此处写法不对，暂未找到
                        List<Object> tempList = sqlSession.selectList(getSqlStatement(SqlMethod.SELECT_BATCH_BY_IDS), idList);
                        updateIdList = tempList.stream().map(e -> tableInfo.getPropertyValue(e, keyProperty)).toList();
                    }

                    // 遍历批数据
                    for (T entity : batchList) {
                        Object id = tableInfo.getPropertyValue(entity, keyProperty);
                        if (StringUtils.checkValNull(id) || !updateIdList.contains(id)) {
                            sqlSession.insert(sqlStatement, entity);
                        } else {
                            MapperMethod.ParamMap<T> param = new MapperMethod.ParamMap<>();
                            param.put(Constants.ENTITY, entity);
                            sqlSession.update(getSqlStatement(SqlMethod.UPDATE_BY_ID), param);
                        }
                    }

                    sqlSession.flushStatements();
                    idxLimit = Math.min(idxLimit + batchSize, size);
                    batchList.clear();
                }
                i++;
            }
        };
        return SqlHelper.executeBatch(getSqlSessionFactory(), log, consumer);
    }

但是！

此处仅需要查询出所有已存在的id，不应该查全量数据，感觉这会浪费很多资源去传输、处理、存储，造成时间延长。能不能加一个仅查询某个字段的方法，业务中也非常实用

[nieqiurong]

String sqlStatement = getSqlStatement(SqlMethod.SELECT_OBJS);
Map<String,Object> params = new HashMap<>();
params.put(Constants.WRAPPER, Wrappers.query().select(keyProperty).in(keyProperty, List.of(1,2,3)));
List<主键类型> objects = sqlSession.selectList(sqlStatement, params);

String sqlStatement = getSqlStatement(SqlMethod.SELECT_LIST);
Map<String,Object> params = new HashMap<>();
params.put(Constants.WRAPPER, Wrappers.query().select(keyProperty).in(keyProperty, List.of(1,2,3)));
List<泛型实体> objects = sqlSession.selectList(sqlStatement, params);