Database management system, method and program for supporting the
mutation of a composite object without read/write and write/write
conflicts
    4.
    发明授权
    Database management system, method and program for supporting the mutation of a composite object without read/write and write/write conflicts 失效
    数据库管理系统,支持复制对象突变的方法和程序,无需读/写和写/写冲突

    公开(公告)号:US5857182A

    公开(公告)日:1999-01-05

    申请号:US786605

    申请日:1997-01-21

    IPC分类号: G06F17/30

    摘要: The system, method, and program of this invention avoids potential write/write conflicts and read/write conflicts when a subcomponent of a composite object (e.g., an ADT) is mutated. The embodiments of this invention define a copy semantic for the mutation function. In one embodiment, a copy function is inserted prior to any mutation function. In a another embodiment, a global compile-time analysis is performed to determine if a write/write or read/write conflict exists; and to eliminate redundant copy constructors if a conflict does exist. In a preferred embodiment, only a local analysis is performed during the parsing phase, thereby avoiding a global compile-time analysis. A mutation safe flag is associated with each parse tree node. A read target leaf parse tree node is set to false while non-leaf parse tree nodes (functions) derive their value from an incoming node, except that constructors and copy constructor functions are always true. Whether or not a copy is made of the composite object (i.e., whether or not a copy constructor is inserted) prior to a mutation is determined according to the setting of the mutation safe flags and according to the following. If a mutation safe flag for a mutation function is false, a copy constructor is inserted for the mutated composite object and the mutation safe flag is set to true. In addition, for update and trigger statements, the mutation safe flag for a mutated target is defaulted to true. Furthermore, related update entries are grouped together and a copy is generated for the common target. The generated copy is used as the common target for all of the mutations caused by the update entries grouped together in order to accumulate all of the desired mutations in a same copy of the composite object.

    摘要翻译: 本发明的系统,方法和程序在复合对象(例如,ADT)的子组件被突变时避免潜在的写入/写入冲突和读/写冲突。 本发明的实施例定义了突变功能的复制语义。 在一个实施例中,在任何突变功能之前插入复制功能。 在另一实施例中,执行全局编译时分析以确定是否存在写/写或读/写冲突; 并且如果存在冲突,则消除冗余的拷贝构造函数。 在优选实施例中,在解析阶段仅执行局部分析,从而避免了全局编译时分析。 突变安全标志与每个解析树节点相关联。 读取目标叶解析树节点设置为false,而非叶解析树节点(函数)从传入节点导出其值,除了构造函数和复制构造函数始终为真。 根据突变安全标志的设置,根据以下来确定复制对象(即,是否插入复制构造函数)之前的副本是否被复制。 如果突变功能的突变安全标志为假,则为突变复合对象插入复制构造函数,并将突变安全标志设置为true。 另外,对于更新和触发语句,突变目标的突变安全标志默认为true。 此外,将相关的更新条目分组在一起,并为共同目标生成副本。 生成的拷贝被用作由更新条目组合在一起引起的所有突变的共同目标,以便累积复合对象的相同副本中的所有期望的突变。

    Optimal storage mechanism for persistent objects in DBMS
    5.
    发明授权
    Optimal storage mechanism for persistent objects in DBMS 失效
    DBMS中持久对象的最佳存储机制

    公开(公告)号:US6065013A

    公开(公告)日:2000-05-16

    申请号:US914394

    申请日:1997-08-19

    IPC分类号: G06F17/30

    摘要: A method, apparatus, and article of manufacture for a computer implemented storage mechanism for persistent objects in a database management system. A statement is executed in a computer. The statement is performed by the computer to manipulate data in a database stored on a data storage device connected to the computer. It is determined that an object is to be stored in an inline buffer. When the object can be entirely stored in the inline buffer, the object is stored in the inline buffer. When the object cannot be entirely stored in the inline buffer, a selected portion of the object is stored in the inline buffer and the remaining portion of the object is stored as a large object.

    摘要翻译: 一种用于数据库管理系统中用于持久对象的计算机实现的存储机制的方法,装置和制品。 在计算机中执行语句。 该语句由计算机执行以操作存储在连接到计算机的数据存储设备上的数据库中的数据。 确定一个对象要存储在一个内联缓冲区中。 当对象可以完全存储在内联缓冲区中时,对象被存储在内联缓冲区中。 当对象不能完全存储在内联缓冲区中时,对象的选定部分被存储在内联缓冲区中,并且对象的剩余部分被存储为大对象。

    Fully integrated architecture for user-defined search
    7.
    发明授权
    Fully integrated architecture for user-defined search 失效
    用于用户定义搜索的完全集成架构

    公开(公告)号:US06278994B1

    公开(公告)日:2001-08-21

    申请号:US09112302

    申请日:1998-07-09

    IPC分类号: G06F1730

    摘要: A method, apparatus, and article of manufacture for a computer-implemented integrated architecture for user-defined search. A statement is executed in a database stored on a data storage device connected to a computer, and the database contains data. Under control of the database, a statement is received that requests manipulation of the data. The statement includes a user-defined function. An index of a user-defined index type is used to access the data. The user-defined function is processed with the accessed data.

    摘要翻译: 用于用户定义搜索的用于计算机实现的集成架构的方法,装置和制品。 在存储在连接到计算机的数据存储设备上的数据库中执行语句,并且数据库包含数据。 在数据库的控制下,接收到要求操纵数据的语句。 该语句包括用户定义的函数。 用户定义的索引类型的索引用于访问数据。 用访问的数据处理用户定义的函数。

    Supporting dynamic tables in SQL query compilers
    9.
    发明授权
    Supporting dynamic tables in SQL query compilers 失效
    在SQL查询编译器中支持动态表

    公开(公告)号:US5930795A

    公开(公告)日:1999-07-27

    申请号:US786603

    申请日:1997-01-21

    IPC分类号: G06F17/30

    CPC分类号: G06F17/30421 G06F17/30474

    摘要: A framework for a query compiler and run-time environment for resolving a table reference to a dynamic table that is first identified at run-time but is initially unknown at compile-time. A parser parses the table reference and creates a parsed representation for the table that identifies the type of dynamic table. A code generator creates executable plans containing run-time table object representations (TAOB), from the parsed representations, that contain the type of dynamic table. The TAOB is also extended to provide for parameters that are definable at run-time, including a table ID of the actual table entity being referenced. A routine is embedded in a run-time routine that checks the type of dynamic table in the TAOB and dispatches control to one of a plurality of corresponding routines, depending upon the type of dynamic table. The corresponding routines use the extended TAOB to dynamically link the table reference to the dynamic table, depending upon the type of dynamic table. Once the dynamic table has been linked to the table reference, query processing continues as if that table were a base or derived table.

    摘要翻译: 用于查询编译器和运行时环境的框架,用于解析对运行时首先标识但最初在编译时未知的动态表的表引用。 解析器解析表引用,并创建用于标识动态表类型的表的解析表示。 代码生成器从解析的表示形式创建包含运行时表对象表示(TAOB)的可执行计划,其中包含动态表的类型。 TAOB还被扩展以提供在运行时可定义的参数,包括被引用的实际表实体的表ID。 例程被嵌入到运行时程序中,该运行时程序检查TAOB中的动态表的类型,并根据动态表的类型将控制调度到多个相应的例程之一。 相应的例程使用扩展TAOB来动态地将表引用动态链接到动态表,这取决于动态表的类型。 一旦将动态表链接到表引用,查询处理将继续进行,就像该表是基类或派生表一样。