1269985 九、發明說明： 【發明所屬之技術領域】 本發明係指一種資料庫系統架構，尤指一種適用於資料庫升級 之多層次資料庫系統架構。 【先前技術】 _ 近年來’隨著多媒體技術、空間資料庫技術和電腦網路的快速 發展，資料庫系統的發展十分迅速，應用領域愈來愈廣，企業單 位、政府部門的行政管理、辦公自動化、企業生産計劃管理、銀 、仃財務管理、醫院病房賴管理、圖書館管理、氣象預報、地理 貢訊系統(GIS)、電子商務等等都離不開資料庫系統。一方面，一 些較成熟的技術，如各種大、中、小和微型電腦資料庫管理系統 和一些傳統的資料庫設計方法已付諸應用層面；另一方面，尚有 鲁許多理論及實際問題亟待開發和探索，如空間資料庫、多媒體資 料庫、網路資料庫、智_料料。制是網路資料庫的理論和 技術問題正成爲資料庫研究的—個熱門研究領域，她之下，傳 統的資料庫技術和系統便顯得力不從心，這也對傳統的技術和研 究開發工作提出了挑戰。 從欠测年代末_來’辭所有的#料縣_發展都以關 聯式貝料庫系統(relatlonal database system)為主轴。事實上關聯 1269985 、、乍法疋5日的主流’而關聯式模型成為資料庫領域中最重要 的發展也是無庸置疑的。簡單的說，關聯式資料庫系統就是： 1. 使用者所看到的資料都是表格。 2. 使用者可使用的運算（例如資料的擷取）都是從舊的表格中 產生新的表格。例如，有些運算子(〇perat〇r)用來榻取表格中的某 些橫列成為-個子集合’另有些運算子則用來選取表格中的直行 成為子市5 Μ然’表格中橫列的子集合及直行的子集合可以 | 合起來構成一個新的表格。 關聯式資料庫之所以如&受歡迎，主要原因是其建立在嚴格的 數學概念基礎Jl〇X「隱運算」為基礎），無論資料庫内的實體 _ 0卿戌實體之間都關係來表示，對資料的檢索結果也是一種 關係(j卩祕）’耻Μ式龍庫_念單…其資料結構也麟 相當簡單、清晰。此外關聯式資料庫中關係表格的存取路徑對使 用者而言是公開的，其具備了更高的資料獨立性、更加的安全保 >在I1 生’亦簡化了資料庫開發建立的工作。總結來說，關聯式資料 表的他點在於對結構化資料關連的定義和控制十分詳細。 然_聯式資料庫的缺闕是缺乏雜’—旦筒_聯發生 異動，則牽-髮而動全身，對應程式的修改是相當f時費力。通 常關聯式資料庫在兩侧立的資料表格(data廳)之外還需要建 立一個_係表格(relational table)，以在資料庫表格中間設^主鍵 (primary key)與其他附屬鍵(foreignkey)的關聯性。請參閱圖。 1269985 知之關聯式資料庫B1示意圖。關聯式資料庫bi包 20:Γ 祕2G，其中絲ω中含樣狂1，表格 -样If4 K2。此外，根據資料K1及資料K2之關係另建立 表才口 30來表不。 建立關係表袼的好處在於搜尋上如果有闕係式的建立，所有相 屬=可以非常快地找出來，·但是相對而言，使用關係表 。、®非节大的缺點就是··當需要新增、删除、修改資料，甚 至需要新增、删除、修改資料庫時，資料之間的關聯性會使得執 行這些任務變得非常複雜、耗_容易出錯。舉例而言，當需要 刪除貧料時，必須先將相_之附屬_資料—刪除後才可以 刪除原本要刪除的資料。如第i圖所示，如果我們要刪除表格2〇 中的資料K2，必須先將表格3〇中所有有關資料κ2的資料刪除才 可以將表格20中的資料Κ2刪除。 從程式設計師的觀點來看，前後資料的變動關係會使得程式設 計變得異常也往往在#料更新的過程中，因為修正程序的 錯誤，而㈣統產生錯誤的訊息，甚至造成整個祕的癱瘓。此 外，因為各資料彼此之__报重要，即使是資料庫的硬體 $些微的差錯’也可㈣失各㈣之咖關聯或破壞了該關聯性 育料庫，造成整儀統停擺。如此—來，軸_性資料庫增加 了資料搜尋上的便繼，卻因此犧牲了射4新增、刪除、修改的 效率，並且增加了程式本身;^要的複雜度；更嚴重的是，複雜 1269985 的關聯式資料庫結構的修改成了一近乎不可能的工程。 針對資料1弟1圖。一個具有三個資料表格的關聯式資料庫在 時，不=身作處理的過程中，若處理的資料涉及鍵值資料_ 4新增、刪除或修改資料，皆需要三個步驟。例如·· (―)新增資料的步驟： V驟―：於表格10新增資料K1 步驟二：於表格20新增資料K2 步驟三··於表格30新增資料K1與資料Κ2之關聯資料 (二）刪除 步驟一： 步驟二： 步驟: 資料的步驟： 刪除表格3〇中資料K1與資料幻之關聯資料 刪除表格1〇中資料Κ1 删除表格20中資料Κ2 (二）修改資料(鍵值資料)的步驟： 步驟— 步驟二 步驟三 修改表格3〇中資料K1與資料尺2之關聯資料 修改表格1〇中之資料K1 修改表袼2〇中之資料K2 上述關於新增、顺與修改資料的步歡有涉及表格π α υ與表 1269985 格1〇、表格20之間的關係，尚未包含表格30與其他表格的關係。 如果Μ式雜庫的具有n㈣料表格，則我們可·期如表格 30的關係表格最多將有q種情形。在表格數量持續增加時，單單 進仃簡單的資料或表格的新增、珊或修改的運算量就十分驚人。 另方面’由於C++等物件導向(object oriented)的程式相繼問 世，幾乎所有程式設計都以物件導向為設計鮮。物件導向設計 春的概念在於將所有最縣喊分析的元件賴祕倾物件的相 行為也因為私式设計的基本元件變成可分析的概念， (imified modeling langUage)語言—類的語言便相繼被發明與強化 來支持物件導向程式的祕。然祿不幸的，目前在程式語言被 •強化成物件導向絲同時，資料庫的觀雖有所謂物件導向的資 •料庫來配合，但是仍為了在麟上與穩定度的考量而採用關聯式 資料庫。 藝 JL因為如此，無論是物件導向的資料庫或關聯式資料庫，當遭 遇系統升級的需求時，資料庫的升級都會遭到無比的風險。總括 而言，這些風險來自： ⑴資料庫表格的攔位往往會因為升級的需求而增加，而在關聯 式資料庫的架構下，「增加攔位」會更動到資料庫基本架構。 (2) 關聯式資料庫的資料表格會因為主表格的變更而跟著被更 動，同樣也會更動到資料庫的基本架構。 (3) 當資料庫版本變更時，資料庫的資料格式可能無法相容。 1269985 由於這些資料庫設計的風險，大部分資訊人員往往視移動關聯 式資料庫的資料為畏途，即使資料庫效能已經不敷使用，也只利 用調教效能(performancetuning)的方式試圖解決。 【發明内容】 因此，本發明之主要目的在提供一種可解決資料庫升級風險之 多層次資料庫系統。 本發明係提供-種多層次的資料庫系统，其包含有一實體資料 層以及-索引資料層。該實體資料層包含一資料表，該資料表記 錄有該資料庫系統之複數筆物件資料，該麟庫祕之資料表之 每筆物件貝料包含複數個貧料項目，該物件資料之該複數個資 料項目之每—資料項目具有一屬性名稱，一物件識別碼，一項目 内容及-索引代碼。該索引資料層記錄有索引資料，用來據以搜 尋資料項目之索引代碼以取得該資料庫系統中符合應用 耘式所要求之物件資料。 本發明另提供-種建立多層次資料庫系統之方法，其下列 步驟··產生一實體資料層，苴包含一 一 資料庫系統之複數筆物件資料，該資料庫系有： 貝科項目具有一屬性名稱物件物，—項目内容及 1269985 一索引代碼；以及依據一索引資料層之索弓丨資料，搜尋該複數個 資料項目之索引代碼以取得料料統中符合應贿式所要求 之物件資料。 【實施方式】 請參考第2圖。第2圖為本發明之多層次資料庫系統1〇〇之示 意圖。多層次貧料庫系統100包含有一素引資料層2〇〇、複數個動 •態貧料層300及一實體資料層400。其中，多層次資料庫系統10〇 係透過索引·層2〇〇與應用程式連結，而索引資料層2〇〇與複 數働態資料層3GG另與實體資料層彻相連結。實體資料層4〇〇 中包含複數個叢集410’而記錄實體資料之資料表43〇則包含於叢 •集410巾。索引資料層200及複數個動態資料層3〇〇中包含複數 .個索引資料表21〇，用來記錄索引資料。索引資料表21〇及其所記 錄的索引資料係依據應用程式之搜尋條件由應用程式產生。索弓丨 馨資料層200及複數個動態資料層3〇〇亦依據應用程式之設計需求 喊生，若應雜式需求較為簡單，多層:欠資料庫系統 100甚至 可以不具備動喊料層綱，·只由索引資料層·及實體資料層 400構成。 §電細系統的應用程式依據搜尋條件及所需求的資料庫類型 產生索1貧料層2〇0之索引資料表210 (及複數個動態資料層300 之索引資料表210)後，應用程式即依據索引資料層2〇〇及動態資 料層300中的索引貢料表210之索引資料來搜尋資料表430。 12 1269985 請參考第3圖。第3圖為本發明之多層次資料庫系統ι〇〇中， 實體資料層400之-叢集410之示意圖。資料表43〇係由物件導 向程式構成，資料表資料表43〇記錄有資料庫系統1〇〇之複 數筆物件資料422，其中每-筆物件資料422包含複數個資料項目 424，物件資料422之複數個資料項目424之每一資料項目具 有一屬性名稱432、一物件識別碼434、一項目内容436及一索引 籲代碼440A。此一實施例之資料表430巾，屬性名稱432、物伊哉 別碼434及項目内容436為多層次資料庫系統1〇〇中的靜態資料， 該靜態資料之部署綱要（Schema)於實體資料層4〇〇建立之初艮 規劃好，當資料庫系統1〇〇運作或進行任何資料處理的動作時， 該部署綱要皆不受影響，亦即麵次資料庫系統⑽巾實體^料 層4〇0之實體資料具有一不需更動的基本架構，可確保實體資料 的穩定性及安全性。如先前所述，實體資料層4〇〇中設置了複數 個叢集410，而為了因應資料庫各種實體資料性質不同，每一叢集 • 410可允許擁有—種部署綱要，屬於該叢集彻之所有資料表43〇 則以同一種部署綱要來設計。 另外，此一實施例之資料表㈣巾，索引代碼44〇A為多層次 資料庫系統100中的動態資料，其係依據索引資料層200及^^ 個動態資料層3〇〇中索引資料表21G的索引資料而建立，而當索 引資料表210有所更動時，索引代碼44〇A亦會同步隨著更動（通 常是重新建立），當索引資料表別被刪除時，其所建立的索引 13 1269985 代碼440A亦隨之删除。 度，資料表43〇中可勺人目應應用程式的設計及資料搜尋的複雜 便是依據*_丨.資^—個以上物丨代碼碰，應用程式 資料項目424之索引件、21G之索引資料’以搜尋資料表430中 符合應用程式所要求^顧’以取綱次#_統1〇0中 目424。 之物件貧料422或物件資料422中的資料項 請繼續參考第4R # 繼為本發明之^庫之-關聯_ 意圖，其中關聯資料二=統100 -實施例之資料表之示 〇之百列顯示攔位名稱44,於關聯資料表 4中，、有記錄識別石馬」、「姓名」、 型」五個攔位。而關胸次上丨士 卞叹血 顯示各筆_魏=^表⑼之其餘各列職據攔位名稱44 時，則將_資料表個之—實施例之舰伽 表)及資料搁位ϋ表名稱（此處命名為「人員」資料 :母聿貝枓5己錄42白具有一唯一的記錄識別 料二二他資料:己錄作區別(主要以該記錄識別碼作為關聯資 :、1、 於只作本發明之一實施例時，則將該記錄識別％ 鱗於資料表伽之第二攔，做為物件識別碼434 一棚的⑽馬 目咖36觸細錄42 _應於資料襴 貝4内谷，以作為項目内容436 —攔的内容。 1269985 以關聯資料表40(「人員」資料表）中第一筆資料記錄们舉 例來5兒’第-筆資料記錄42記錄有「姓名」攔位内容：「陳建宏」、 .性別」攔位内谷·「男」、「年齡」攔位内容：「3〇」、「血 型」攔位内容：「A」，另外第一筆資料記錄42的記錄識別碼為 「1」’則本發明之-實施例的實作方式會據此建立一筆物件資料 422，其包含四筆資料項目424，其中第一筆資料項目424的屬性 名稱432 -欄的資料内容為「人員•姓名」（如前所述，該棚内 • 容為關聯資料表40中資料表名稱與資料攔位名稱的合併），物件 識別碼434 一攔的内容為Γ1」，項目内容伽一搁的内容為「陳 建f」；同理，第二筆資料項目424的屬性名稱432 一攔的資料 内容為「人貞•性別」，物件識別碼434 -攔的内容為Γι」（皆 •屬^一物件資料，因此物件識別碼相同），項目内容436 -攔的 '内合則為「男」…餘此類推。如此一來，關聯資料表40中之一筆 =料記錄42會建立成本發明之—實施例之資料表之四筆項目 資料此四筆項目資料424構成一筆物件資料似，因此四 項目資料424的物件識別碼434 -攔的内容皆相同。1269985 Nine, the invention description: [Technical field of the invention] The present invention refers to a database system architecture, especially a multi-level database system architecture suitable for database upgrade. [Previous technology] _ In recent years, with the rapid development of multimedia technology, spatial database technology and computer network, the development of the database system is very rapid, the application field is more and more extensive, the administrative management and office of enterprise units and government departments Automation, enterprise production planning management, silver, financial management, hospital ward management, library management, weather forecasting, geographic gambling system (GIS), e-commerce, etc. are all inseparable from the database system. On the one hand, some mature technologies, such as various large, medium, small and micro computer database management systems and some traditional database design methods have been applied. On the other hand, there are still many theoretical and practical problems. Development and exploration, such as spatial database, multimedia database, network database, wisdom materials. The theory and technical issues of the network database are becoming a hot research field in the research of the database. Under her, the traditional database technology and system are not enough, which also puts forward the traditional technology and research and development work. challenge. From the end of the year of the under-test, all the #县县_ developments are based on the relatingal database system. In fact, the correlation between 1269985 and the mainstream of the 5th day, and the related model has become the most important development in the field of database is unquestionable. To put it simply, the associated database system is: 1. The information that the user sees is a table. 2. The operations available to the user (such as data retrieval) are generated from the old table. For example, some operators (〇perat〇r) are used to take some of the columns in the table into a sub-collection. Other operators are used to select the straight rows in the table to become sub-markets. The sub-collections and the straight sub-collections can be combined to form a new table. The reason why the relational database is as popular as & is that it is based on the strict mathematical concept basis Jl〇X "hidden operation", regardless of the relationship between the entities in the database. It is said that the search result of the data is also a kind of relationship (j 卩 secret) 'shame-style dragon _ _ sing... its data structure is also quite simple and clear. In addition, the access path of the relational table in the relational database is open to the user, and it has higher data independence and more security. In I1, it also simplifies the work of database development. . In summary, the point of the associated data sheet is that the definition and control of the association of structured data is very detailed. However, the lack of a joint database is a lack of miscellaneous ─ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Usually, the relational database needs to establish a relational table in addition to the data table (data hall) on both sides to set the primary key and other auxiliary keys in the middle of the database table. Relevance. Please refer to the figure. 1269985 Schematic diagram of the associated database B1. Correlation database bi package 20: 秘 Secret 2G, where silk ω contains madness 1, form - like If4 K2. In addition, according to the relationship between the data K1 and the data K2, another form is established 30 to indicate. The advantage of establishing a relational table is that if there is a shackle in the search, all dependencies = can be found very quickly, but relatively speaking, use the relational table. The disadvantage of the non-large section is that when you need to add, delete, modify data, or even need to add, delete, or modify the database, the correlation between the data will make the implementation of these tasks very complicated and consuming. Error-prone. For example, when it is necessary to delete the poor material, the data to be deleted must be deleted before deleting the attached data. As shown in Figure i, if we want to delete the data K2 in Form 2〇, we must delete all the data in Table 3〇 related to κ2 before deleting the data Κ2 in Form 20. From the point of view of the programmer, the change of the data before and after will make the programming become abnormal and often in the process of updating the material, because the error of the program is corrected, and (4) the wrong message is generated, and even the whole secret is caused. paralysis. In addition, because each data is important to each other, even the hardware of the database is slightly inferior. In this way, the Axis-Sex database increases the efficiency of the data search, but it sacrifices the efficiency of the addition, deletion, and modification of the shot 4, and increases the complexity of the program itself; ^more importantly; The modification of the associative database structure of Complex 1269985 became a nearly impossible project. For the data 1 brother 1 map. An associated database with three data tables is not in the process of processing. If the processed data involves key data _ 4 adding, deleting or modifying data, it takes three steps. For example, (-) Steps for adding data: V--: Adding information K1 in Form 10 Step 2: Adding information K2 in Form 20 Step 3 · Adding information related to K1 and Data 于2 in Form 30 (2) Deletion Step 1: Step 2: Steps: Data Steps: Delete the data in the table 3〇 and the information in the illusion. Delete the data in the table 1〇 删除1 Delete the data in the table Κ2 (2) Modify the data (key value Steps of the data): Steps - Step 2 Step 3: Modify the related information in Table 3, Data K1 and Data Ruler 2 Modify the data in Form 1〇 K1 Modify the data in Table 2〇 K2 The above information about new, smooth and modified The step of the data involves the relationship between the table π α υ and the table 1269998, the table 20, and the relationship between the table 30 and other tables is not included. If the 杂-type miscellaneous library has n (four) material tables, then we can have as many as q cases. As the number of forms continues to increase, the amount of new, old, or modified calculations for simple data or forms is amazing. On the other hand, since object oriented programs such as C++ have been introduced one after another, almost all programming is designed with object orientation. The concept of object-oriented design spring lies in the phase behavior of all the most screaming components of the county, and the basic elements of the private design become the analyzable concept. (imified modeling langUage) language-like language has been successively Invention and reinforcement to support the secret of the object-oriented program. Unfortunately, at present, the programming language is strengthened into object-oriented silk. At the same time, although the database view has a so-called object-oriented resource library, it is still used in order to consider the stability of the column. database. Because of this, whether it is an object-oriented database or an associated database, when the system is upgraded, the database upgrade will be incomparable. In summary, these risks come from: (1) The barriers to the database tables tend to increase due to the need for upgrades. Under the framework of the associated database, the “additional barriers” will change to the basic structure of the database. (2) The data table of the associated database will be changed as the main form changes, and the basic structure of the database will be changed. (3) When the database version is changed, the data format of the database may not be compatible. 1269985 Due to the risk of designing these databases, most information personnel often regard the data of the mobile-related database as a fearful way. Even if the database performance is insufficient, it can only be solved by means of performance tuning. SUMMARY OF THE INVENTION Accordingly, it is a primary object of the present invention to provide a multi-level database system that addresses the risk of database upgrades. The present invention provides a multi-level database system comprising a physical data layer and an index data layer. The entity data layer includes a data table, and the data table records the plurality of object data of the database system. Each item of the data item of the library data includes a plurality of poor materials items, and the plurality of items of the item data Each of the data items has a property name, an object identification code, a project content, and an index code. The index data layer records index data for searching the index code of the data item to obtain the object data required by the application system in the database system. The invention further provides a method for establishing a multi-level database system, the following steps: generating an entity data layer, comprising a plurality of object data of a database system, the database is: Attribute name object, project content and 1269985 an index code; and searching for the index code of the plurality of data items according to the data of an index data layer to obtain the object data in the material system that meets the requirements of the bribe type . [Embodiment] Please refer to Figure 2. Figure 2 is a schematic representation of the multi-level database system of the present invention. The multi-level lean repository system 100 includes a prime data layer 2, a plurality of dynamic lean layers 300, and a physical data layer 400. The multi-level database system 10 is linked to the application through the index layer 2, and the index data layer 2〇〇 and the complex data layer 3GG are further connected to the entity data layer. The entity data layer 4〇〇 contains a plurality of clusters 410' and the data sheet 43 of the recorded entity data is included in the clusters 410. The index data layer 200 and the plurality of dynamic data layers 3〇〇 include a plurality of index data tables 21〇 for recording index data. The index data table 21 and the index data recorded by it are generated by the application based on the search criteria of the application. The cable material layer 200 and the plurality of dynamic data layers 3 are also called according to the design requirements of the application. If the demand for the hybrid type is relatively simple, the multi-layer: the database system 100 may not even have the layer of the shouting layer. Only consists of the index data layer and the entity data layer 400. § The application of the electric system generates the index data table 210 (and the index data table 210 of the plurality of dynamic data layers 300) according to the search condition and the type of the required database, and the application program The data table 430 is searched according to the index data layer 2 and the index data of the index metric table 210 in the dynamic data layer 300. 12 1269985 Please refer to Figure 3. FIG. 3 is a schematic diagram of the cluster 410 of the entity data layer 400 in the multi-level database system ι of the present invention. The data table 43 is composed of an object-oriented program, and the data sheet data table 43 records the plurality of object data 422 of the database system 1 , wherein each-object data 422 includes a plurality of data items 424, and the object data 422 Each of the plurality of data items 424 has an attribute name 432, an item identification code 434, an item content 436, and an index call code 440A. The data table 430 of this embodiment, the attribute name 432, the object identification code 434, and the item content 436 are static data in the multi-level database system, and the deployment schema (Schema) of the static data is in the entity data. The layer 4〇〇 is initially planned, and when the database system operates or performs any data processing, the deployment outline is not affected, that is, the face database system (10) towel entity layer 4实体0's physical data has a basic structure that does not require changes to ensure the stability and security of physical data. As described earlier, a plurality of clusters 410 are set in the entity data layer 4, and in order to respond to the different nature of the various entities of the database, each cluster 410 can allow possession of a deployment outline, belonging to all the data of the cluster. Table 43 is designed with the same deployment outline. In addition, in the data table (four) of this embodiment, the index code 44A is the dynamic data in the multi-level database system 100, which is based on the index data layer 200 and the dynamic data layer 3 in the index data table. The index data of 21G is established, and when the index data table 210 is changed, the index code 44〇A is also synchronized with the change (usually re-establishment), and the index established when the index data table is deleted. 13 1269985 Code 440A is also deleted. Degrees, data sheet 43 can be used in the design of the application and the complexity of the data search is based on *_丨. 资 ^ more than one object code, application data item 424 index, 21G index data 'In the search data table 430, it meets the requirements of the application program ^ to take the outline #_统1〇0中目424. For the items in the object poor material 422 or the object data 422, please refer to the 4R # 继 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The column displays the block name 44, and in the related data table 4, there are five blocks for the record identification "stone horse", "name", and type. On the other hand, when the sighs of the sighs are displayed on the chest, the rest of the _Wei=^ table (9) are listed as 44, and the _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ϋ table name (here named "person" information: mother 聿 枓 枓 5 has recorded 42 white has a unique record identification material 22 two data: has been recorded as a difference (mainly with the record identification code as a related asset: 1. When only one embodiment of the present invention is used, the record identifies the % scale in the second block of the data table gamma, as the object identification code 434, and the shed (10) the horse eye 36 touches the record 42 _ The information is as follows: 襕 内 内 内 1 1 1 1 1 1 1 1 1 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 269 "Name" block content: "Chen Jianhong", "Gender" Blocking Neigu, "Men", "Age" block content: "3〇", "Blood Type" block content: "A", the first pen The record identification code of the data record 42 is "1"', and the implementation of the embodiment of the present invention will establish an object accordingly. The material 422 includes four data items 424, wherein the information content of the attribute name 432-column of the first data item 424 is “person/name” (as mentioned above, the inside of the shed is included in the related data table 40) The combination of the name of the data table and the name of the data block), the content of the object identification code 434 is Γ1", the content of the item content is "Chen Jianf"; similarly, the attribute name of the second data item 424 The content of the 432 block is "People's Sex and Gender", and the object identification code 434 - the content of the block is Γι" (all of them belong to an object data, so the object identification code is the same), and the item content 436 - the block's internal Then, it is a "male"... and so on. In this case, one of the records in the associated data table 40 will create a four-item project data of the invention--the data sheet of the embodiment. The information is similar, so the object identification code 434-block of the four item data 424 has the same content.
睛參考第5 ®、第6圖。如前所述，應用程式依據搜尋條件及 應用程式所需求的資料庫類型而產生索引資料層2〇〇之索弓I資 表21〇(及複數個動態資料層3〇〇之索引資料表21〇後），、I 紐搜尋。第5圖、第6圖說明於料資5 建立有-索引資料表210，其中包含索引資料用來將 424的項目内容436中具有資料内容為「男」的項目資料似貝於 15 1269985 »亥員目資料424之索引代碼.A棚位中標示 項目資料424的項目内容436中具有資料為〇 ’並用來將 /:，於該項目資料424之索引代碼4·二 當索引資料表21〇建立完成後，會於 =1。 的資料表動態增加一索引代碼4雜=^ 二二:，目資料⑽ 之項目貧料424之索引代碼衡一搁則不作用。子口條件 當欲依據索5丨資料表加所標示时㈣料來搜尋 · t付^的物件資料422時，例如欲搜尋資料庫中所有男性員工的 資料，則只需要於索引代碼偷中提取出代碼内容為、'〇,貝^ 料項目似，亚依據資料項目424中物件識別碼4 件資料422。如此完成了資料的搜尋動作。 '斤而的物 同理，如欲找出資料表430中，年齡小於30歲之男性員工， 貝懷據事先於索㈣料層建立的索引f料表，或於動態資 料層300中建立的索引資料表21〇)另於資料表指建立之一索引 代碼440B ’再與原先已建立之索引代碼衡交叉搜尋即可提 取出兩個條件皆符合的資料。 此外’配合應用程式需求而欲達成「於習知的關聯資料表仙 中新增-攔位」的目的時，於本發明之—實施_資料表的 作法則是直接於資料表430中新增-筆#料項目424,其各搁位内 16 1269985 容=前=建立方式建立。例如要新增一 「婚姻狀態」的搁位’ 、筆資料。己錄42的相對於「婚姻狀態」攔位的攔位内容為厂已 婚j ’則於資料身a 、十表430令新增一筆資料項目424，其屬性名稱 432 —攔的資料内容為「 、—、 谷局人貝•婚姻狀態」，物件識別碼434 —攔 的内谷為1」（屬同一物件資料，因此物件識別碼相同），項目 内容436 -攔的内容則為「已婚」。 明參考第7圖。第7圖為本發明之多層次資料庫系統100欲刪 Γ物件雜的作法示賴。若配合剌程式需求而欲達成「於 j的關％貝料表4G中刪除—筆f料記錄42」甚或「於習知的關 P貝料表4G中刪除_攔位」的目的時，於本發明之—實施例的資 ;表的作法則是直接於索引f料層之索引資料表,中， rnmttmm m fm 43〇^mm 422 、目24的索引貝料表21()標示出有效或常用的索引資料。 2〇〇 -可依需求將處理資料的單御見劃為以資料項目似 ^ ^ 422 . 430 ^^ 0 t 以上的索引資料表210 ’以實現同時針對叢集彻、 細2及資料項目424等實體資料範圍作全域 索引的味。當應用程式欲刪除資料表中第三筆物件 · 時’於全域索引資料表21〇中將「可用性」攔位標示為「、否 則「全域蚓」攔位對應資料表之「有效索引」之索弓丨」 17 1269985 亦將被解頃為無效的資料。 一 U ±i rir 可藉由於索引次伽_ ^，多層次貢料庫系統100即 姻中有\二θ巾建立*⑽料細標示岭體資料層 項目424 表BO、叢集MO、物件資料422或資料 層料表430之「有效索引」亦為索引資料 曰7 W耻式4相建立之_索引欄，其與物件識別碼 不目同，無論是物件識別碼434或有效索引（甚至是其他動能 胃_式峨求峨嫩引或局料 引0 ' 第8〜圖為本發明多層次資料庫系、统ι〇〇另一實施例中實體資料 層叢集貫作方法的示意圖。#資料_因為經年累月的發展與改 文而不斷增加時，貫體資料層4⑻的資料表獨並不需要無限制 的擴張。本發明之多層次資料庫系統刚於實體資料層另建 立叢集410以放置複數個資料| 43〇。叢集41〇本身被系統設定有 固定的容量關，崎絲-叢集410的大小無限纖大。當一 類的物件資料數量增長至某-程度時，整個資料庫會重新開啟一 個叢集彻，以放置新增的資料表携及其中的物件資料。在一個 大型的資料庫系統巾’若是能對搜尋的資料再作進一步的分類與 篩選，則有助於搜尋效率的提升，因此，於建立叢集41〇時，本 發明即對叢集410進行建立索引的動作，並於索引資料層2〇〇中 建立一搜哥叢集索引的對照表220以將搜尋對象作進一步的範圍 18 1269985 限制。而倾資料層中複數個叢集41G_f被限定儲存在 單-電腦域巾。複數個錢41〇分聊轉於複數倾由網路系 統相連之電腦主機中，以分散主機系統的處理負載。 為配合使用者仍較習慣以表格方式建立資料庫的資料内容，應 用程式的介面仍舊可以設計成喊格（或_表格）的形式供使 用者輸入資料’無論是資料的新增、刪除、修改，或是資料搁位 的新增、刪除、修改，磁於-做上表格之_龍關係的建 立移除，皆可透過上述索引資料層200或動態資料層於實 體貝料層働的貝料表43〇中建立相關的索引代碼柳棚位（一 =或-個以上）以完成物件資料422的建立。以前述實施例而言， ΓΓ者欲建立—筆具有廿侧位㈣料時，習知作法就是於資 二中寫人—筆具有廿侧⑽資料，而本發明之錢次資料庫 2⑽之—實施觸是於資録中建立廿筆具有三個攔位 尚未建立索引代碼的情況下）的資料項目424。為避免於 貝⑽項目424建立時佔用系統資源以致於影響多層次資料庫系統 的政能，於應用程式端與多層次資料庫系統1〇〇之索引資料層 二間可另外建立-仔列緩衝區，使多層次資料庫系統刚可利 用所有系關置的_來進行資料的建立工作。 使用另私方^ ’ _搜尋之後的資料也可賤触歧計以轉換成 ^習慣的表格方式呈現，對於朗程式而言，多層次資料 〜100提供-安全的資料儲存架構，而使用者端可以沿用原 1269985 有處理方式以避免使__改變。也由於㈣次資料庫 ^改mf轉物件胃料的方錢乎是錢的，針對資料庫的架 US 轉求，多料㈣庫祕卿仰完成這些 對關#式貝料庫系統而言相當具風險的變更。 、斤述僅為本發明之較佳實施例，凡依 處均等__侧她^職目 【圖式簡單說明】 第1圖為習知關聯式資料庫帽係表格之示意圖。 第頂為本發明多層次資料庫系統—實施例之示意圖。 第3圖為本發3移層次資料庫纽之資料表結構示意圖。 第4圖為本發明多層次資料庫系統中建立—資料表之示音圖。 第5圖為本發明多層次資料庫系統中—索引資料表之示音圖。 第6圖為本發❹層錢料輕統巾於資料表中建立糾之示意 圖0 第-7圖為本發明多層次資料庫系統中刪除_物件資料之矛音圖。 第8圖為本發明多層次資料庫系統另—實施例中實體資料層叢集 實作方法的示意圖。 20 1269985 【主要元件符號說明】 ία、K2 表格貧料 Β1 關聯式資料庫 10、20、30 表格 40 關聯資料表 42 資料記錄 44 欄位名稱 50 應用程式 100 多層次資料庫系 統 200 索引資料層 210 索引資料表 220 對照表 300 動態資料層 400 實體資料層 410 叢集 422 物件資料 424 資料項目 430 資料表 432 屬性名稱 434 物件識別碼 436 項目内容 440,440A，440B 索引代碼 21Refer to the 5th and 6th figures. As mentioned above, the application generates the index data layer 2 according to the search conditions and the type of database required by the application. 2) The index data table 21 of the plurality of dynamic data layers 3 After the )), I search for New Zealand. Figure 5 and Figure 6 illustrate the establishment of the index data table 210 in the material 5, which contains the index data used to set the project content of the 424 item content 436 with the data content of "male" as if it were 15 1269985 » The index code of the member data 424. The item content 436 of the item data 424 in the booth of the shed has the data 〇' and is used to set /:, in the index code of the item data 424, the index data table 21 is established. When done, it will be =1. The data table dynamically increases an index code 4 miscellaneous = ^ 22:, the target data (10) of the project poor material 424 index code balance does not work. When the sub-port condition is to be searched for the object data 422 of the t-subscription according to the data sheet (4), for example, if you want to search for the data of all the male employees in the database, you only need to extract the index code. The content of the code is, '〇, the item is similar to the material item, and the item identification code 4 is 422 according to the object identification code 424. This completes the search for data. Similarly, if you want to find out the male staff of the age less than 30 years old in the data table 430, Becker is based on the index f material table established in advance (4), or established in the dynamic data layer 300. Index data table 21〇) In addition, the data sheet refers to the establishment of one of the index codes 440B' and then cross-search with the original index code to extract the data that meets both conditions. In addition, in order to achieve the purpose of "additional-blocking in the associated data table" in conjunction with the application requirements, the method of implementing the data sheet of the present invention is directly added to the data table 430. - Pen # material item 424, within each of the shelves 16 1269985 capacity = before = establishment mode established. For example, you need to add a "marital status" position and pen data. The content of the block 42 relative to the "marriage status" block is factory married j', and a new data item 424 is added to the data body a, ten table 430 order, and the attribute name 432 - the data content of the block is " , —, 谷局人贝 • Marriage status, object identification code 434 — the inner valley of the barrier is 1” (the same object data, so the object identification code is the same), the content of the item 436 - the content of the block is "married" . See Figure 7 for details. Figure 7 is a representation of the multi-level database system 100 of the present invention for deleting objects. If the purpose of the program is to meet the purpose of "removing the _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ In the embodiment of the present invention, the method of the table is directly in the index data table of the index f layer, rnmttmm m fm 43〇^mm 422, and the index of the item 24 of the item 24 () is marked as valid or Commonly used index data. 2〇〇- According to the requirements, the single-view information of the processing data can be classified into the index data table 210 of the data item like ^^ 422 . 430 ^^ 0 t or more to achieve simultaneous clustering, detailing 2 and data items 424, etc. The scope of the entity data is the taste of the global index. When the application wants to delete the third object in the data table, the 'Availability' block is marked as "" in the global index data table 21", otherwise the "Global Index" block corresponds to the "Active Index" of the corresponding data table. The bow 丨 17 17269985 will also be dissolved as invalid information. A U ± i rir can be attributed to the index sub-gamma ^ ^, multi-level tribute library system 100, that is, there is a \ 2 θ towel establishment * (10) material fine labeling ridge data layer project 424 table BO, cluster MO, object data 422 Or the "effective index" of the data layer table 430 is also the index column of the index data 曰 7 W shame 4 phase, which is not the same as the object identification code, whether it is the object identification code 434 or the effective index (or even other Kinetic stomach _ 峨 峨 峨 峨 或 或 或 局 局 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' Because of the continuous development and rewriting of texts, the data sheet of the data layer 4(8) does not need unrestricted expansion. The multi-level database system of the present invention just creates a cluster 410 in the entity data layer to place a plurality of Data | 43〇. The cluster 41〇 itself is set by the system with a fixed capacity, and the size of the Saki-Plug 410 is infinitely large. When the number of objects in a class increases to a certain level, the entire database will reopen a cluster. Thorough to place new The data sheet carries the information about the objects in it. In a large database system, if you can further classify and filter the searched data, it will help the search efficiency increase. Therefore, when the cluster is established, 41 The present invention is an action of indexing the cluster 410, and a lookup table 220 of the Sogou cluster index is created in the index data layer 2 to limit the search object to a further range of 18 1269985. The cluster 41G_f is limited to be stored in the single-computer domain towel. The plurality of money 41〇 is divided into multiple computer hosts connected to the network system to distribute the processing load of the host system. The form of the database is used to create the data content of the database. The interface of the application can still be designed in the form of a shuffle (or _ form) for the user to input data 'whether it is new, deleted, modified, or new data. Adding, deleting, modifying, and magnetically-making the _Dragon relationship of the form to be removed, all through the above index data layer 200 or dynamic data layer in the physical material The relevant index code will be established in the 贝 料 table 43〇 ( more than one) to complete the establishment of the object data 422. In the foregoing embodiment, the latter wants to establish - the pen has the 廿 side position (four) In the case of the material, the conventional practice is to write the person in the second capital. The pen has the side (10) information, and the money database 2 (10) of the present invention is implemented in the record. The pen has three blocks and has not yet been indexed. In the case of the code, the data item 424. In order to avoid the system resources occupied by the Bayes (10) project 424 to affect the multi-level database system, the index data layer of the application side and the multi-level database system can be additionally established. District, so that the multi-level database system can just use the _ of all departments to set up the data. Using the other private ^ ' _ search data can also be used to convert the data into a custom form. For the Lang program, the multi-level data ~ 100 provides - a secure data storage architecture, and the user side The original 1269985 can be used to avoid __ change. Also because of the (four) times database to change the mf to transfer the material of the stomach material is money, for the database of the US transfer, more material (four) library secrets to complete these pairs of closed-type library system is quite Risky changes. The description of the present invention is only a preferred embodiment of the present invention, and is based on the equalization of the __ side of her position [Simplified description of the drawings] Figure 1 is a schematic diagram of a conventional associated database hat form. The top is a schematic diagram of an embodiment of the multi-level database system of the present invention. The third figure is a schematic diagram of the structure of the data table of the 3-shift hierarchical database. Figure 4 is a diagram showing the establishment of a data sheet in the multi-level database system of the present invention. Figure 5 is a sound diagram of the index data table in the multi-level database system of the present invention. Fig. 6 is a schematic diagram of the establishment of a correction for the light-weight towel of the hair layer in the data sheet. Figure 0-7 shows the spear sound map of the deleted object data in the multi-level database system of the present invention. Figure 8 is a schematic diagram showing the implementation method of the entity data layer cluster in another embodiment of the multi-level database system of the present invention. 20 1269985 [Description of main component symbols] ία, K2 Forms 贫1 Related Databases 10, 20, 30 Table 40 Related Data Table 42 Data Record 44 Field Name 50 Application 100 Multi-level Database System 200 Index Data Layer 210 Index data table 220 comparison table 300 dynamic data layer 400 entity data layer 410 cluster 422 object data 424 data item 430 data table 432 attribute name 434 object identification code 436 item content 440, 440A, 440B index code 21