RU2389066C2 - Multidimensional database and method of managing multidimensional database - Google Patents

Abstract

FIELD: information technology.

SUBSTANCE: invention relates to ordered information arrays logically organised into databases stored in physical carriers, e.g. magnetic discs, and to methods of managing databases realised in form of software tools which process information using computers. The method of managing databases involves the following operations. Information is identifying by comparing actual data coordinate values with unified (sorted into types according to recording format) values recorded in the database. The obtained set of standard coordinate values defines the set of cells of address pointers and recording blocks for the identified data. Recording, reading, alteration, deleting and compression of information is carried out through multidimensional space cell referencing and subsequent switching to the data block according to the address pointer. In case of index structure application, multidimensional space cell referencing is carried out by bypassing the binary index tree.

EFFECT: reduced database exchange and time for accessing the database.

16 cl, 4 dwg

Description

FIELD OF KNOWLEDGE

The invention relates to ordered arrays of information logically organized into databases stored on physical media, such as magnetic disks, and to database management methods implemented in the form of software tools that process information using computers.

The database (hereinafter referred to as the DB) is characterized by its logical data model and consists of separate records - data containing distinctive characteristics. Data groups are grouped into subsets.

The database information includes the actual data and metadata identifying the data in accordance with the systematics of the subject area. Information in the database is also divided into structured - with records of the same type of data in a single format, and unstructured - with records of data of mismatched types in various formats.

The database management method is based on the operations of identification, writing, reading, changing and deleting information from the database, compressing information in the database, controlling the internal consistency of information in the database during the joint work of two or more users, restoring the internal consistency of information after the emergency operation of the database.

There are several logical models of database data that differ in their internal structure and control methods:

- hierarchical (Hierarchical DB), the main feature of this model is the hierarchical grouping of individual data within the database, such that each data of the lower level is associated with only one data of the higher level of the hierarchy. Such a model imposes significant restrictions on the speed of reading or writing data in the case of using access logic other than a given hierarchical structure. An example of a hierarchical model is IMS (IBM);

- Network (Network DB), is the development of a hierarchical model. Each data in the database can be associated with one or more data of a higher level. In the network model, there are no restrictions on the speed of reading or writing data when changing the access logic, while the operations of changing or deleting data require a long time to reorganize numerous data links. An example of a network model is Adabas (SAP);

- Relational (Relation DB), consisting of many two-dimensional arrays - relational tables containing homogeneous data characteristics, combined into subsets and interconnected by logical relationships. Spacing the characteristics of relational tables ensures the independence of the external representation of data from the organization of the internal structure of the database. At the same time, the time for reading, writing, changing or deleting data is significantly increased due to the need for bulk operations to separate / restore data to / from individual characteristics. An example of a relational model is Oracle (Oracle);

- multidimensional (Multidimensional DB), logically represented as a multidimensional space defined by coordinate axes - dimensions (data types), hypercubes - areas of multidimensional space (subsets of data) and vectors - cells of multidimensional space (data). Organization of access to information in the form of direct addressing to cells of multidimensional space provides high speed of all types of data operations. However, with an increase in the number of measurements and their coordinates, the volume of the database increases many times (exponentially), while most of the cells in the multidimensional space remain blank with meaningful data, and the speed of access to the database decreases sharply. An example of a multidimensional model is Analysis Services (Microsoft).

BACKGROUND AND PROTOTYPE

The subject of the invention relates to multidimensional databases and methods for managing them. As a prototype of the selected invention, described in US patent No. 5359724 from October 25, 1994

The difference between the logical data model of the prototype and other multidimensional databases is based on the use of two structural levels in the prototype model, designed to store data and coordinates of multidimensional space, respectively.

At the lower structural level contains a complete set of all possible combinations of coordinates, which are cells of multidimensional space. The cells have address pointers to data blocks located at the upper structural level. Each data block is defined for a limited set of coordinates in order to aggregate predominantly significant data in blocks.

The number of cells in multidimensional space significantly exceeds the number of data blocks. The cells in which the address pointers are located can be indexed and combined into a binary tree type structure.

The structured and unstructured information is stored in the database in separate files interconnected by logical links.

The database control method includes the following operations.

Information is identified by comparing the actual values of the data coordinates with their unified (typed by recording format) values previously recorded in the database. The resulting set of reference coordinates determines the choice of location cells for address pointers and recording blocks of identified data.

Writing, reading, changing, deleting and compressing information is carried out by accessing a multidimensional space cell and then navigating to the data block at the address pointer. In the case of applying the index structure, access to the cell of multidimensional space is carried out by traversing the binary index tree.

Information is recorded in the existing data block or, in its absence, in a new block after its formation together with the address pointer, multidimensional space cell index and the binary tree branch.

Reading information is carried out after finding the desired block and enumerating the data values in the block.

Changing or deleting information is performed using the appropriate adjustment of the data in the block, the reorganization of the structure of the blocks and balancing the branches of the binary tree in the case of indexing the cells of multidimensional space.

Information is compressed in the database by removing data blocks for which there are no address pointers in the cells of multidimensional space.

Monitoring the internal consistency of information in the database during the joint work of two or more users is carried out by sequentially blocking their access to the data used.

The restoration of internal information consistency after an emergency shutdown of a database is performed by the method of forward writing of data to an additional file - a transaction log. The transaction log is limited and requires constant updating by replacing old data with new data. When the database is launched, the latest time data recorded in the database files and the transaction log are compared. If there is no data stored in the transaction log file in the database file, they are transferred to the first file. Otherwise, the most recently recorded data is excluded from the database file.

The prototype of the invention allows to reduce the volume of the database due to the organization of storage at the second structural level of predominantly significant data. However, the presence at the first structural level of a complete set of cells of multidimensional space leads to an increase in the volume of the database and access time to it.

The technical result of the present invention is to reduce the volume of the database and access time to it.

SUMMARY OF THE INVENTION

The technical result is achieved by forming a database (DB), the logical model of which includes:

- measurements, each of which determines the type of data and is represented by standardized coordinate values and codes of their identifiers;

- hypercubes, each of which defines a subset of data and is represented by coordinate identifiers associated with address pointers with identifiers of data vectors with matching codes written to the file last;

- data vectors, each of which is represented by coordinate identifiers associated with address pointers with data vector identifiers with matching codes previously written to the file, while the coordinates of the data vector coordinates written to the file first are connected by address pointers with hypercube coordinate identifiers with matching codes.

The coordinate identifier consists of the following set of elements:

- the name of the identifier, consisting of the magnitude of its offset relative to the beginning of the file and the serial number of the entries in the vector file;

- an identifier code consisting of a digital measurement code and a digital code of a uniform coordinate value in this measurement;

- a set of address pointers.

Address pointers come with four types of pointers.

The address pointer of the first type consists of the offset value relative to the beginning of the file of the same type address pointer of the previous coordinate identifier with the same code and the serial number of this coordinate identifier in a cyclic list formed by address pointers of identifiers of coordinates of data vectors and a hypercube.

The address pointer of the second type consists of the offset from the beginning of the file of the same type pointer of the previous coordinate identifier and the sequence number of the coordinate identifier in the cyclic list formed by the address pointers of the coordinate identifiers that are part of one vector of this data, while the first coordinate order in the record is associated with the last in the order of writing by the identifier of the coordinate.

The address pointer of the third type consists of the offset value relative to the beginning of the file of the same type address pointer of the previous coordinate identifier at a higher hierarchy level and the sequence number of the hierarchy level of this coordinate identifier, while the address identifiers of coordinate identifiers that are part of the same hierarchical group vectors of a given form a finite list.

The address pointer of the fourth type consists of the offset value relative to the beginning of the file of the same type address pointer identifier of the coordinate included in one variant group, or, in the case of the completion of one variant group, which is part of another variant group, and the sequence number of the variant group, with address identifiers of identifiers coordinates that are part of one vector of this, form a finite list.

A hypercube consists of coordinate identifiers, each of which is represented as a descriptor vector. The address pointer of the first type of the descriptor vector contains the offset value relative to the beginning of the file and the serial number of the address pointer of the first type of identifier of the coordinate of the vector given with the matching identifier code recorded in the file last.

To order information, for example, in chronological order, the descriptor vectors in the hypercube can be logically combined into two or more segments.

The composition of each segment of the hypercube includes a root vector consisting of one identifier of the coordinate with the system code.

The root vector and descriptor vectors that are part of one hypercube segment are connected into the index structure of a binary tree type by address pointers of coordinate identifiers of the second and third types.

Root vectors and descriptor vectors with matching coordinate identifier codes included in different segments of the same hypercube are connected into circular lists by address pointers of coordinates identifiers of the fourth type.

The root vectors of all database hypercubes are connected in a circular list by address pointers of identifiers of coordinates of the first type, the sequence numbers of which take the value of the numbers of hypercubes in the order they are written to the file.

To provide access for several users at the same time to the database file recorded, for example, on magnetic media, a permanent information array is placed in the computer’s RAM - a buffer.

The buffer contains the following elements:

- tables of values of descriptor vectors with the largest sequence numbers of address pointers of coordinates of the fourth type;

- a list of tables of descriptor vector values;

- inverted lists of coordinates of data vectors;

- data vectors involved in reading and writing information.

The list of descriptor vector value tables contains the addresses of the location of the tables in the buffer. The list of tables is formed in the order in which the tables are written to the file.

Each table of descriptor vector values corresponds to one of the dimensions of the logical data model of the database.

Rows contain sets of fields, each of which corresponds to one of the hypercubes formed by measurements corresponding to tables.

One set contains at least the following fields:

- field of the magnitude of the offset name of the identifier of the vector of the descriptor relative to the beginning of the file

- field of the serial number of the address pointer of the fourth type of identifier of the coordinate of the descriptor vector,

- the field of the magnitude of the offset relative to the beginning of the file of the address pointer of the first type identifier of the coordinate of the vector of the data with the matching code recorded in the file last, and

- the field of the serial number of the address pointer of the first type of identifier of the coordinate of this vector given.

In addition to these sets of fields, the rows of the tables may additionally contain fields of unified coordinate values, fields of indexes of unified coordinate values, and address fields of inverted coordinate lists of data vectors in the buffer.

For storing structured (data vectors, descriptor vectors, root vectors) and unstructured information (coordinate values, text comments, graphic images, audio, video, and other records) in a common file, data vectors under certain conditions may contain fragments of unstructured information .

The technical result in terms of a method of accelerating data access and database management achieve the following features of operations:

- information identification is performed by digitally encoding standardized coordinate values as part of measurements and replacing actual coordinate values by digital codes as part of data vectors;

- writing, reading, changing, deleting and compressing information is ensured through the formation of logical sections of multidimensional space, represented by cyclic lists of identifiers of coordinates of data vectors and hypercubes with matching identifier codes;

- control of the internal consistency of information during the joint work of two or more users is ensured by sequential numbering of data vectors in the order they are written to a file;

- restoration of the internal consistency of information after the emergency shutdown of a multidimensional database is provided by means of the forward record of coordinate identifiers in the composition of descriptor vectors relative to the moment of their recording in the composition of data vectors.

Identification of information in the database is based on a comparison of the actual coordinate values of a given data received from an external source of information with standardized coordinate values as part of measurements previously stored in the database.

In case of discrepancy between the actual and unified coordinate values, the actual values are included in the measurements as new unified values with the assignment of digital identifier codes to them.

This identification is completed by replacing the coordinate values with the digital codes of their identifiers.

Information is recorded in the database by separate data vectors after forming a linked list of coordinate identifiers, including fragments of unstructured information, writing new descriptor vectors to the file, or changing the values of descriptor vectors previously written to the file.

Reading information from the database is carried out on the basis of a search key compiled by the user from one or more identifiers of coordinates using five step procedures.

Changing information in the database is carried out at a logical level by writing several versions of this vector.

Information is removed from the database at the logical level by writing several versions of the given vector.

The subsequent reading of the deleted information is carried out by forming a consolidated version of this vector, consisting of coordinate identifiers with nonzero values of the coordinates identifier codes and fragments of unstructured information enclosed between them.

Information is compressed in the database at the physical level by saving only consolidated versions of data vectors in a file. For the entire database volume or for individual hypercubes, the following actions are performed.

The internal consistency of information in the database is controlled by the joint work of two or more users with the same data by performing the operations of changing or deleting information in the sequence by which users issue commands to write changed (deleted) versions of data vectors.

The restoration of internal information consistency after an abnormal shutdown of a database is carried out by comparing the values of the descriptor vectors and the data vectors written to the file between them before the crash.

DESCRIPTION OF GRAPHIC MATERIAL

Figure 1. The scheme of address relationships of a given vector with one hierarchical and two variant groups of coordinate identifiers in its composition.

Figure 2. The scheme of address relationships of vectors of descriptors of a hypercube with two segments.

Figure 3. The block diagram of the buffer.

Figure 4. The scheme of information retrieval in the database.

DETAILED DESCRIPTION OF THE INVENTION

DESCRIPTION OF THE DEVICE

The technical result of the present invention is to reduce the volume of the database and access time to it.

The technical result in terms of the device is achieved by the formation of a logical database data model, including:

- measurements, each of which determines the type of data and is represented by standardized coordinate values and codes of their identifiers;

- hypercubes, each of which defines a subset of data and is represented by coordinate identifiers associated with address pointers with identifiers of data vectors with matching codes written to the file last;

- data vectors, each of which is represented by coordinate identifiers associated with address pointers with data vector identifiers with matching codes previously written to the file, while the coordinates of data vector coordinates written to the file first are connected by address pointers with hypercube coordinate identifiers with matching codes.

The database structure is a horizontal scan of a multidimensional space filled with only significant data. Their coordinates can be part of hierarchical and variant groups as part of the same vector.

The coordinate identifier consists of the following set of elements:

- the name of the identifier, consisting of the magnitude of its offset relative to the beginning of the file and the serial number of the entries in the vector file;

- an identifier code consisting of a digital measurement code and a digital code of a uniform coordinate value in this measurement;

- a set of address pointers.

Address pointers come with four types of pointers.

The address pointer of the first type consists of the offset value relative to the beginning of the file of the same type pointer pointer of the previous coordinate identifier with the same code and the serial number of this coordinate identifier in a cyclic list formed by address pointers of identifiers of coordinates of data vectors and a hypercube.

The address pointer of the second type consists of the offset from the beginning of the file of the same address pointer of the previous coordinate identifier and the sequence number of the coordinate identifier in the cyclic list formed by the address pointers of the coordinate identifiers that are part of one vector of this data, while the first coordinate order in the record is associated with the last in the order of writing by the identifier of the coordinate.

The address pointer of the third type consists of the offset value relative to the beginning of the file of the same type address pointer of the previous coordinate identifier at a higher hierarchy level and the sequence number of the hierarchy level of this coordinate identifier, while the address identifiers of coordinate identifiers that are part of the same hierarchical group vectors of a given form a finite list.

The address pointer of the fourth type consists of the offset value relative to the beginning of the file of the same type of the address identifier of the coordinate identifier included in one variant group, or, in the case of the completion of one variant group included in another variant group, and the sequence number of the variant group, with address identifiers of identifiers coordinates that are part of one vector of this, form a finite list.

An example of a vector of this, consisting of three identifiers of coordinates that are part of one hierarchical and two variant groups is shown in figure 1. Numbers 1, 2, 3, 4 - indicate address pointers of the corresponding types.

A hypercube consists of coordinate identifiers, each of which is represented as a descriptor vector. The address pointer of the first type of the descriptor vector contains the offset value relative to the beginning of the file and the serial number of the address pointer of the first type of identifier of the coordinate of the vector given with the matching identifier code recorded in the file last.

In order to streamline information, for example, in chronological order, descriptor vectors in a hypercube can be logically combined into two or more segments.

The structure of each segment of the hypercube includes a root vector consisting of one identifier of the coordinate with the system code.

The root vector and descriptor vectors, which are part of one segment of the hypercube, are connected to the index structure of a binary tree type by address pointers of coordinate identifiers of the second and third types.

The root vectors and descriptor vectors with matching coordinate identifier codes included in different segments of the same hypercube are linked into circular lists by address pointers of coordinates identifiers of the fourth type. In this case, the ordinal numbers of the address identifiers of the coordinate identifiers are assigned the values of the numbers of segments of the hypercube in the order they are written to the file.

The root vectors of all database hypercubes are connected into a cyclic list by address pointers of identifiers of coordinates of the first type, the serial numbers of which are assigned the values of the numbers of hypercubes in the order they are written to the file.

A diagram of the address relationships of the hypercube descriptor vectors with two segments is shown in FIG. Numbers 1, 2, 3, 4 - indicate address pointers of the corresponding types.

In order to provide users with access to the database file recorded, for example, on a magnetic disk, on an additional (or alternative) storage medium, for example, in the computer’s RAM, a constant information array (temporary data set) is located - a buffer.

The buffer contains the following elements:

- tables of values of descriptor vectors with the largest sequence numbers of address pointers of coordinates of the fourth type;

- a list of tables of descriptor vector values;

- inverted lists of coordinates of data vectors;

- data vectors involved in reading and writing information.

The list of descriptor vector value tables contains the addresses of the location of the tables in the buffer. The list of tables is formed in the order in which the tables are written to the file.

Each table of descriptor vector values corresponds to one of the dimensions of the logical data model of the database. Rows of tables correspond to the coordinates of the measurements and are formed in the order of writing standardized coordinate values in a file.

Rows contain sets of fields, each of which corresponds to one of the hypercubes formed by measurements corresponding to tables.

One set includes:

- the field of the magnitude of the offset name of the identifier of the vector of the descriptor relative to the beginning of the file,

- field of the serial number of the address pointer of the fourth type of identifier of the coordinate of the descriptor vector,

- the field of the magnitude of the offset relative to the beginning of the file of the address pointer of the first type identifier of the coordinate of the vector of the data with the matching code recorded in the file last, and

- the field of the serial number of the address pointer of the first type of identifier of the coordinate of this vector given.

In addition to these sets of fields, the rows of the tables may additionally contain fields of unified coordinate values, fields of indexes of unified coordinate values, and address fields of inverted coordinate lists of data vectors in the buffer.

Inverted lists of coordinates of data vectors contain the addresses of the data vectors in the buffer.

Each vector given in the buffer is assigned a sequence number equal to the largest serial number of the entries in the file of one of the identifiers of the coordinates of the vector of this. The sequence number of this vector is used to provide multi-user access to the database.

The block diagram of the buffer is shown in Fig.3.

For storing structured (data vectors, descriptor vectors, root vectors) and unstructured information (coordinate values, text comments, graphic images, audio, video, and other records) in a common file, the data vectors can contain fragments of unstructured information when the following conditions:

- a fragment of unstructured information as part of a given vector begins and ends with coordinate identifiers;

- coordinate identifiers reflect the data type and recording format of fragments of unstructured information.

The technical result in terms of the device also achieve the following features of operations:

- information identification is provided by digitally encoding standardized coordinate values as part of measurements and replacing actual coordinate values by digital codes as part of data vectors;

- writing, reading, changing, deleting and compressing information is ensured through the formation of logical sections of multidimensional space, represented by cyclic lists of identifiers of coordinates of data vectors and hypercubes with matching identifier codes;

- control of the internal consistency of information during the joint work of two or more users is ensured by sequential numbering of data vectors in the order they are written to a file;

- restoration of the internal consistency of information after the multidimensional database crashes, is ensured by the forward recording of coordinate identifiers in the composition of descriptor vectors relative to the moment of their recording in the composition of data vectors.

Identification of information in the database is based on a comparison of the actual coordinate values of a given data received from an external source of information with standardized coordinate values as part of measurements previously stored in the database.

In case of discrepancy between the actual and unified coordinate values, the actual values are placed in the measurement as new unified values with the assignment of digital identifier codes to them.

Information was recorded in the database in batches equal to a separate data vector after forming a linked list of coordinate identifiers, including fragments of unstructured information, writing new descriptor vectors to the file, or changing the values of descriptor vectors previously written to the file.

The offset value in the name of the coordinate identifier is calculated as the sum of the values of the previously filled file volume and the offset of the name of this coordinate identifier relative to the beginning of the given vector. The sequence number of the record of the vector of this to the file is generated as a unique integer increasing in value.

The offset value of the address pointer of the first type of identifier of the coordinate of the vector of this is determined based on the current value of the vector of the descriptor with the matching identifier. The initial values of the descriptor vector at the time of recording the first identifier of the coordinate of the vector of this have a link to the location address in the descriptor vector file in terms of the offset value of the address pointer of the first type and a single value in terms of the serial number of this address pointer.

The offset value of the address pointer of the second type of identifier of the coordinate of the vector of the given is set as the sum of the values of the previously filled file volume and the offset of the same type of address pointer of the previous identifier of the coordinate relative to the beginning of the vector of this. For the first identifier in the composition of the vector of this second term is the offset value of the address pointer of the last identifier relative to the beginning of the vector of this. The serial number of the address pointer is set by the position of the identifier of the coordinate in the vector of this.

The offset value of the address pointer of the third type of the identifier of the coordinate of the vector of this is given as the sum of the values of the previously filled file volume and the offset relative to the beginning of the vector of the same type address pointer of the previous identifier of the coordinate located at a higher level of the hierarchy. For the first identifier of the coordinate in the composition of the vector of this second term is the offset value of its own address pointer. The sequence number of the address pointer is given by the hierarchy level of the coordinate identifier.

The offset value of the address pointer of the fourth type of identifier of the coordinate of the vector of this given as the sum of the values of the previously filled file volume and the offset relative to the beginning of the vector of the same type of the address pointer of the previous identifier of the coordinate included in this variant group, or, if completed, the offset relative to the beginning of the vector of the same type of address pointer previous triad, which is part of another variant group. For the first identifier of the coordinate in the composition of the vector of this second term is the offset value of its own address pointer. The serial number of the address pointer is given taking into account the identity of the coordinate identifier in this variant group.

In the case of the presence of one or more fragments of unstructured information in the vector, this fragment size is taken into account when calculating the displacement values of the coordinate identifiers following it.

If there are no descriptor vectors with coordinate identifiers in the hypercube that coincide with the identifiers of the coordinates of the newly formed vector of this, the specified descriptor vectors remain written to the file.

The calculated values of the address pointers of the identifiers of coordinates of the first type are recorded in a file at the location of the descriptor vectors. This vector is written to a new place at the end of the file (after writing to the descriptor vectors the calculated values of the address pointers of the coordinate identifiers of the first type).

To search and read information from the database, a search key is used, made up of one or more coordinate identifiers.

DESCRIPTION OF THE METHOD OF OPERATION OF THE DEVICE

The technical result in terms of the method of accelerating access to data is achieved by the database management method, which is characterized by the following operations:

- preliminary identification of information by digitally encoding standardized coordinate values as part of measurements and replacing actual coordinate values by digital codes as part of data vectors;

- writing, reading, changing, deleting and compressing information by forming logical sections of multidimensional space represented by cyclic lists of identifiers of coordinates of data vectors and hypercubes with matching identifier codes;

- control of the internal consistency of information during the joint work of two or more users by sequential numbering of data vectors in the order they are written to a file;

- restoration of the internal consistency of information after the emergency shutdown of a multidimensional database is provided by means of the forward record of coordinate identifiers in the composition of descriptor vectors relative to the moment of their recording in the composition of data vectors.

Identification of information in the database is performed by comparing the actual values of the coordinates of a given data received from an external source of information with standardized values of coordinates in the composition of measurements previously stored in the database.

In case of discrepancy between the actual and unified coordinate values, the actual values are included in the measurements as new unified values with the assignment of digital identifier codes to them.

This identification is completed by replacing the coordinate values with the digital codes of their identifiers.

Information is recorded in the database in batches equal to a separate data vector after forming a linked list of coordinate identifiers, including fragments of unstructured information, writing new descriptor vectors to the file, or changing the values of descriptor vectors previously written to the file.

The offset value in the name of the coordinate identifier is calculated as the sum of the values of the previously filled file volume and the offset of the name of this coordinate identifier relative to the beginning of the given vector. The sequence number of the record of the vector of this file is generated in the form of a unique integer increasing in value.

The offset value of the address pointer of the first type of identifier of the coordinate of the vector of this is determined based on the current value of the vector of the descriptor with the matching identifier. The initial values of the descriptor vector at the time of recording the first identifier of the coordinate of the given vector contain a link to the location address in the descriptor vector file in terms of the offset value of the address pointer of the first type and a single value in terms of the serial number of this address pointer.

The offset value of the address pointer of the second type of identifier of the coordinate of the vector of the data is determined as the sum of the values of the previously filled file volume and the offset of the same type of address pointer of the previous identifier of the coordinate relative to the beginning of the vector of this. For the first identifier in the composition of the vector of this second term is the offset value of the address pointer of the last identifier relative to the beginning of the vector of this. The serial number of the address pointer is set by the value corresponding to the position of the identifier of the coordinate as part of this vector.

The offset value of the address pointer of the third type of identifier of the coordinate of the vector of this data is determined as the sum of the values of the previously filled file volume and the offset relative to the beginning of the vector of the same type address pointer of the previous identifier of the coordinate located at a higher level of the hierarchy. For the first coordinate identifier in the vector of this second term is the offset of its own address pointer. The serial number of the address pointer is set by the value of the hierarchy level of the coordinate identifier.

The offset value of the address pointer of the fourth type of identifier of the coordinate of the vector of this data is determined as the sum of the values of the previously filled file volume and the offset relative to the beginning of the vector of the same type address pointer of the previous identifier of the coordinate included in this variant group, or, if completed, the offset relative to the beginning of the vector of the same type of address pointer previous triad, which is part of another variant group. For the first identifier of the coordinate in the composition of the vector of this second term is the offset value of its own address pointer. The serial number of the address pointer is set based on the identity of the coordinate identifier for this variant group.

If one or more fragments of unstructured information are included in the vector, the fragment size is taken into account when calculating the offset values of the coordinate identifiers following it.

If there are no descriptor vectors with coordinates identifiers in the composition of the hypercube that coincide with the identifiers of the coordinates of the newly formed vector of this, these descriptor vectors are written to the file.

The calculated values of the address pointers of the identifiers of coordinates of the first type are recorded in a file at the location of the descriptor vectors. This vector is written to a new place at the end of the file after writing to the descriptor vectors the calculated values of the address pointers of the coordinate identifiers of the first type.

Reading information from the database is carried out on the basis of a search key compiled by the user from one or more coordinate identifiers.

At the first step in the buffer, in the list of addresses of tables of descriptor vectors, the addresses of tables whose positions in the list correspond to digital measurement codes are searched. After the transition to the tables selected as a result of the search, they are searched for in their composition of rows whose positions correspond to digital codes of standardized coordinate values.

If there is no value of at least one of the desired descriptor vectors, the reading is completed by generating a negative search report.

If the buffer contains a complete set of values of the descriptor vectors, the sequence numbers of the address pointers of the fourth type of identifiers of the coordinates of the descriptor vectors are compared. If they are equal to the sequence number corresponding to the last segment of the hypercube, the values of the found descriptor vectors are compared with each other by the value of the sequence numbers of the address pointers of the first type, followed by the choice of the descriptor vector with the lowest number. Otherwise, a transition is made to the hypercube segment stored in the database file.

At the second step, they search for descriptor vectors in the hypercube segment by traversing the binary tree of the index structure. In case of equality of the sequence numbers of the address pointers of the fourth type, the descriptor vectors are compared with each other by the value of the sequence numbers of the address pointers of the first type, followed by the selection of the descriptor vector with the lowest number. Otherwise, the transition to the next segment of the hypercube is carried out with the repetition of the second step until the sequence numbers of the address pointers of the fourth type of descriptor vectors coincide.

In the third step, based on the offset value of the address pointer of the first type of the descriptor vector, the identifier of the coordinate of the vector of the given one with the lowest serial number is read. In accordance with the address pointers of the second type, the remaining identifiers of the coordinates of the vector of this are read and compared with the identifiers of the coordinates of the search key.

If they coincide, one of the matching identifiers of the coordinates of the vector of the given one with the largest serial number of the address pointer of the second type is determined. Otherwise, we proceed to the fifth step of the current sequence of reading information.

At the fourth step, hierarchical and variant groups of the vector of the given are bypassed, starting with the matching coordinate identifier with the highest serial number of the address pointer of the second type. If the identifiers of the coordinates of the search key are included in the same hierarchical and variant groups of a given vector, it is included in the search report.

At the fifth step, as part of the read vector of this, a comparison is made between the number of address numbers of the address pointers of the first type of coordinate identifiers corresponding to the search key, the selection of the coordinate identifier with the smallest sequence number, reading the coordinate identifier of the previous vector given by the address pointer of the first type and reading other identifiers of coordinates of the previous vector given by address pointers of the second type.

The fourth and fifth steps are repeated in the order of approaching the beginning of the file until you read one of the identifiers of the coordinates of the vector of the data corresponding to the search key, with the serial number of the address pointer of the first type equal to one.

Changing information in the database is carried out at a logical level by writing several versions of the vector given in the following sequence.

At the first step, the current version of the given vector is read.

At the second step, an altered version of the vector of the given one is composed of one or more identifiers of coordinates and / or fragments of unstructured information changed relative to the current version, and one invariable identifier of the coordinate.

In the third step, a modified version of the vector of the data is recorded, with the values of the offset relative to the beginning of the file recorded in the first version of the vector of the data stored in the names of the coordinate identifiers.

Subsequent reading of the changed information is carried out by forming a consolidated version of this vector, consisting of one coordinate identifier with a constant identifier code and one or more last-in-order records of the changed coordinate identifiers and / or fragments of unstructured information.

Information is removed from the database at the logical level by writing several versions of the vector given in the following sequence:

at the first step, the current version of the vector of the given is read;

at the second step, a remote version of the vector of the given is formed as a part of one or several identifiers of coordinates removed from the current version of the vector of the given and one constant identifier of the coordinate;

in the third step, a remote version of the vector is recorded, with the codes of remote coordinate identifiers replaced with zero values.

The subsequent reading of the deleted information is carried out by forming a consolidated version of this vector, consisting of coordinate identifiers with nonzero values of the coordinates identifier codes and fragments of unstructured information enclosed between them.

Information is compressed in the database at the physical level by saving only consolidated versions of data vectors in a file. For the entire database volume or for individual hypercubes, the following actions are performed.

At the first step, a second instance of the hypercube is created, associated with the first instance by address pointers of the fourth type of root vectors and descriptor vectors. The second instance of the hypercube is used to record newly received information from an external source.

In the second step, a copy of the first instance of the hypercube is created that is not associated with the first instance of the hypercube with address pointers of the fourth type of root vectors and descriptor vectors. A copy is a receiver of information previously recorded in a hypercube, the first copy is a source of information.

In the third step, information is compressed by forming consolidated versions of the data vectors contained in the information source and transferring them to the information receiver.

In the fourth step, the information source is replaced by the information receiver by linking the second instance and a copy of the first hypercube instance with address pointers of the fourth type of root vectors and descriptor vectors. The first instance of the hypercube is deleted from the file.

The internal consistency of information in the database is controlled by the joint work of two or more users with the same data by performing operations of changing or deleting information in the sequence of issuing instructions by users to write changed (deleted) versions of data vectors.

To ensure the indicated sequence, an information array is formed in the buffer from consolidated versions of data vectors participating in the operations of reading or writing information. The consolidated version of the vector of the data in the buffer is assigned the sequence number of the record equal to the largest sequence number of the record in the file of one of the identifiers of the coordinates of the vector of this.

In the first step, each user, independently of the others, reads information that begins with a search in the buffer for a consolidated version of a given vector corresponding to the search key. If this version of the vector is not in the buffer, this is read from the file.

At the second step, users independently create modified (deleted) versions of the given vector, which are temporarily assigned the record number of the previously read consolidated version of the given vector.

At the third step, by the command to record the user who first completed the formation of the changed (deleted) version of the vector of the data, a second reading is performed in the buffer or file of the consolidated version of the vector of this. Execution of write commands for other users is blocked.

In the fourth step, a comparison is made between the sequence numbers of the records of the modified (deleted) version and the re-read consolidated version of this vector. If the sequence numbers of the records coincide, a new consolidated version of the vector of the data with the current serial number of the record is formed in the buffer, taking into account the first user changing or deleting information. The modified (deleted) version of the vector of the data with the same current serial number of the record is written to the file.

If the sequence numbers of the records of the modified (deleted) version and the re-read consolidated version of the vector of the data do not match, the modified (deleted) and consolidated versions with the current serial number of the record are again formed. The new consolidated version replaces the previous consolidated version in the buffer. The new modified (deleted) version is written to the file. Execution of write commands is unlocked by repeating the third and fourth steps for each subsequent user.

In the fifth step, the last consolidated version of this vector is removed from the buffer in the absence of read or write commands for a specified period of time.

The restoration of internal information consistency after an emergency shutdown of a database is carried out by comparing the values of the descriptor vectors and the data vectors written to the file before the emergency shutdown.

At the first step, based on the comparison of the values of the vectors of the descriptors of the hypercube, they move to the identifier of the coordinate of the last vector of the given one, the most offset from the beginning of the file, while using address pointers of the second type, go to other identifiers of coordinates of the last vector of this.

If the last vector of this complete set of coordinate identifiers is present, the restoration of internal information consistency after an emergency shutdown is completed.

Otherwise, carry out the following actions.

At the second step, the values of the descriptor vectors, the identifier codes of which correspond to the identifiers of the coordinates of the previous data vectors, are replaced by the offset values relative to the beginning of the file and the serial numbers of the address pointers of the first type of identifiers of the coordinates of the vector given.

At the third step, the transition to the last identifier of the coordinate of the penultimate vector of the given offset from the first identifier of the coordinate of the last vector of this data by the amount specified by the format of the identifier of the coordinate is carried out. After that, the address pointers of the second type make the transition to the remaining identifiers of coordinates that are part of the penultimate vector of this.

At the fourth step, the codes of the identifiers of the coordinates of the penultimate vector of the given are compared with the codes of the identifiers of coordinates of the vectors of the descriptors corresponding to the codes of the identifiers of coordinates that are absent from the last vector of this.

When the codes coincide, the values of the descriptor vectors are replaced by the offset values and serial numbers of the address pointers of the first type of coordinates identifiers of the penultimate vector of this.

If the codes do not coincide, they move to the following in the order of approaching the beginning of the file the data vectors until the identifiers of coordinates are found whose codes coincide with the codes of identifiers of coordinates that are not present in the last vector of this data, followed by the replacement of the values of the descriptor vectors.

In the fifth step, the last vector of the given incomplete set of coordinate identifiers is deleted from the file. The restoration of the internal consistency of information after the emergency shutdown of the database is completed.

EXAMPLE OF IMPLEMENTATION

As an example implementation of the present invention consider the database of subscribers of telephone numbers.

The database data schema includes the following measurements (their digital codes are indicated in brackets):

- unique identification numbers (UIN) of subscribers (1);

- last names of subscribers (2);

- names of subscribers (3);

- names of regions (4);

- names of cities (5);

- street names (6);

- house numbers (7);

- UIN phones (8);

- telephone codes of regions (9);

- telephone codes of cities (10);

- subscriber phone numbers (11);

- UIN of telephone codes (12).

The multidimensional database space includes a hypercube of subscribers, a hypercube of telephones, and a hypercube of telephone codes. The constant coordinate of the first hypercube is the subscriber’s UIN, the second hypercube is the telephone UIN, and the third hypercube is the telephone code UIN.

The subscriber’s hypercube data consists of the following coordinates, united in a hierarchy:

1 level hierarchy - subscriber UIN

2 hierarchy level - last name of the subscriber

3 hierarchy level - subscriber name

2 hierarchy level - region

3 level hierarchy - city

4th level hierarchy - street

5 level hierarchy - house number

2 level hierarchy - UIN phone.

The phone’s hypercube data consists of the following coordinates, united in a hierarchy:

Hierarchy Level 1 - Phone UIN

2 hierarchy level - region telephone code

3 hierarchy level - telephone code of the city

4 hierarchy level - subscriber phone number.

The phone code hypercube data consists of the following coordinates, united in a hierarchy:

Level 1 hierarchy - UIN phone code

2 hierarchy level - region telephone code

3 hierarchy level - region

3 hierarchy level - telephone code of the city

4th level of the hierarchy is the city.

The coordinates of this vector can be included in several variant groups. For example, a subscriber may have different addresses and more than one phone number.

The physical record of information in a database file consists of individual numbers logically divided into vectors, coordinate identifiers and their constituent elements using a given recording format.

In a conditional record of the structure of consolidated versions of data vectors, the rows correspond to the identifiers of the coordinates of vectors. The first sign “.” Separates the serial numbers of the hierarchical and variant groups of components in the vector, the second sign “.” - digital codes for the measurement and the unified coordinate value, the third sign “.” - the symbolic names of the measurement and coordinate values conventionally given in the example. The “/” sign separates serial numbers, numeric codes, and symbolic names.

The conditional notation of descriptor vectors and unconsolidated versions of data vectors in the database structure consists of separate paragraphs — vectors, lines — identifiers of coordinates and numbers — constituent elements separated by the “/” sign. Each hypercube in the example consists of one segment.

STRUCTURE OF CONSOLIDATED VERSIONS OF DATA VECTORS.

Hypercube subscribers

Consolidated Version of Vector

1.1 / 1.1 / PIN of the subscriber. one

2.1 / 2.1 / Surname. Smirnov

3.1 / 3.1 / Name. Ivan

2.1 / 4.1 / Region. Tatarstan

3.1 / 5.1 / City. Kazan

4.1 / 6.1 / Street. Flower

5.1 / 7.1 / House number. 128

2.1 / 8.1 / UIN phone. one

2.2 / 4.2 / Region. Karelia

3.2 / 5.2 / City. Petrozavodsk

4.2 / 6.2 / Street. Lake

5.2 / 7.2 / House number. 67

2.2 / 8.2 / UIN phone. 2

Hypercube Phones

Consolidated version of the first vector

1.1 / 8.1 / UIN phone. 1>

2.1 / 9.1 / Telephone code of the region. 560

3.1 / 10.1 / Telephone code of the city. 39

4.1 / 11.1 / Subscriber phone number. 20847

Consolidated version of the second vector

1.1 / 8.2 / UIN phone. 2

2.1 / 9.2 / Telephone code of the region. 315

3.1 / 10.2 / Telephone code of the city. 76

4.1 / 11.2 / Subscriber phone number. 18428

Hypercube phone codes

Consolidated version of the first vector

1.1 / 12.1 / PIN code. one

2.1 / 9.1 / Telephone code of the region. 560

3.1 / 4.1 / Region. Tatarstan

3.1 /10.1 / Telephone code of the city. 39

4.1 / 5.1 / City. Kazan.

Consolidated version of the second vector

1.1 / 12.2 / PIN code. 2

2.1 / 9.2 / Telephone code of the region. 315

3.1 / 4.2 / Region. Karelia

3.1 / 10.2 / Telephone code of the city. 76

4.1 / 5.2 / City. Petrozavodsk.

The structure of the root vector of a hypercube segment (using the example of the root vector of a hypercube of subscribers).

ID Name

000 - the offset value relative to the beginning of the identifier file of the coordinate of the root vector of the descriptor (in relative or absolute selected arbitrary units).

01 - serial number of the entry of the root vector of the descriptor to the file.

Address pointer of the first type.

283 - the offset value relative to the beginning of the file of the address pointer of the first type of identifier of the coordinate of the last in order of recording the root vector of the descriptor in the database.

01 - serial number of the hypercube in the database.

Address pointer of the second type

032 - offset value relative to the beginning of the address pointer file of the second type of the fifth descriptor vector located at the first level of the nodes of the binary tree

04 - node level of the binary tree on which the root vector of the descriptor is located.

Address pointer of the third type.

033 - the offset value relative to the beginning of the address pointer file of the third type of the fifth descriptor vector located at the first level of the nodes of the binary tree.

04 - node level of the binary tree on which the root vector of the descriptor is located.

The address pointer of the fourth type.

004 - the offset value relative to the beginning of the file of the address pointer of the fourth type of identifier of the coordinate of the root vector of the descriptor.

01 - serial number of the segment.

Identifier.

000 is the system digital code for the hypercube segment.

01 - system digital code of the root vector of the descriptor.

The structure of the hypercube descriptor vector (for example, the first subscriber hypercube descriptor vector).

The name of the identifier.

006 - offset value relative to the beginning of the identifier file of the coordinate of the first descriptor vector (in arbitrary units).

02 - serial number of the descriptor vector record to the file.

Address pointer of the first type.

271 - offset value relative to the beginning of the file of the address pointer of the first type of identifier of the coordinate of the last vector of the given with the digital identifier code 001.01

04 - the number of data vectors with digital identifier code 001.01

Address pointer of the second type.

032 - offset value relative to the beginning of the address pointer file of the second type of the fifth descriptor vector located at the first level of the nodes of the binary tree

04 - node level of the binary tree on which the first vector of the descriptor is located.

Address pointer of the third type.

033 - the offset value relative to the beginning of the address pointer file of the third type of the fifth descriptor vector located at the first level of the nodes of the binary tree.

04 - node level of the binary tree on which the root vector of the descriptor is located.

The address pointer of the fourth type.

010 - the offset value relative to the beginning of the file of the descriptor vector of the hypercube with a matching identifier.

01 - serial number of the hypercube segment. Identifier

001 - digital measurement code "subscriber UIN".

01 - digital code of a unified coordinate value.

The structure of this vector (for example, the first identifier of the coordinates of the first version of the first vector of this subscriber’s hypercube).

ID Name

048 - the offset value relative to the beginning of the file of the first identifier of the coordinate of the vector of the given (in arbitrary units).

09 - the serial number of the recording vector given in the file.

Address pointer of the first type.

007 - the offset value relative to the beginning of the file of the address pointer of the first type of identifier of the coordinates of the vector descriptor of the hypercube subscribers.

01 - serial number of the first identifier of the first vector given in the cyclic list of coordinate identifiers with digital code 001.01

Address pointer of the second type.

086 - the offset value relative to the beginning of the file of the address pointer of the second type of the last identifier of the coordinate of the vector of this.

01 - serial number of the first identifier of the coordinate in the vector of this.

Address pointer of the third type.

051 - the offset value relative to the beginning of the file of the address pointer of the third type of the last identifier of the coordinate of the vector of this.

01 - serial number of the coordinate hierarchy level.

The fourth type of address pointer

052 - the offset value relative to the beginning of the file of the address pointer of the fourth type of the first identifier of the coordinate of the vector of this.

01 - serial number of the variant group of coordinates.

Identifier

001 - digital measurement code "subscriber UIN".

01 - digital code of the uniform value of the coordinate "Smirnov".

DB structure Hypercube subscribers measuring Root vector 000.01 / 283.01 / 032.04 / 033.04 / 004.01 / 000.01 First descriptor vector 006.02 / 271.04 / 032.04 / 033.04 / 010.01 / 001.01 Subscriber's UIN Second descriptor vector 012.03 / 055.01 / 032.04 / 033.04 / 018.01 / 002.01 surname Third descriptor vector 018.04 / 061.01 / 002.03 / 045.03 / 022.01 / 003.01 name Fourth Descriptor Vector 024.05 / 067.01 / 014.03 / 147.03 / 028.01 / 004.01 region Fifth descriptor vector 030.06 / 073.01 /164.01 /171.01 / 034.01 / 005.01 city Sixth descriptor vector 036.07 / 079.01 / 008.03 /183.03 / 040.01 / 006.01 Street Seventh Descriptor Vector 042.08 / 085.01 / 032.04 / 033.04 / 046.01 / 007.01 house The first version of the first vector of this 048.09 / 007.01 / 086.01 / 051.01 / 052.01 / 001.01 Subscriber's UIN 054.09 / 013.01 / 050.02 / 051.02 / 052.01 / 002.01 surname 060.09 / 019.01 / 056.03 / 057.03 / 058.01 / 003.01 name 066.09 / 025.01 / 062.04 / 063.02 / 064.01 / 004.01 region 072.09 / 031.01 / 068.05 / 069.03 / 070.01 / 005.01 city 078.09 / 037.01 / 074.06 / 075.04 / 076.01 / 006.01 Street 084.09 / 043.01 / 080.07 / 081.05 / 082.01 / 007.01 house Hypercube Phones Root vector 090.10 / 001.02 / 104.04 / 105.04 / 094.01 / 000.01 First descriptor vector 096.11 / 121.01 / 092.03 / 225.03 / 100.01 / 008.01 Win phone Second descriptor vector 102.12 / 127.01 / 218.01 / 231.01 /106.01 / 009.01 region telephone code Third descriptor vector 108.13 / 133.01 / 104.04 / 105.04 / 112.01 / 010.01 area code Fourth Descriptor Vector 114.14 / 139.01 / 104.04 / 105.04 / 118.01 / 011.01 phone number The first version of the first vector of this 120.14 / 097.01 / 140.01 / 123.01 / 124.01 / 008.01 Win phone 126.14 / 103.01 / 122.02 / 123.02 / 124.01 / 009.01 region telephone code 132.14 / 109.01 / 128.03 / 129.03 / 130.01 / 010.01 area code 138.14 / 115.01 / 134.04 / 135.04 / 136.01 / 011.01 phone number Hypercube subscribers Eighth Descriptor Vector 144.15 / 157.01 / 032.04 / 033.04 / 148.01 / 008.01 Win phone The second version of the first vector of this 048.16 / 043.02 / 158.01 / 153.01 / 154.01 / 001.01 Subscriber's UIN 156.16 / 133.01 / 152.02 / 153.02 / 154.01 / 008.01 Win phone Ninth Descriptor Vector 162.17 / 193.01 / 020.02 / 039.02 /166.01 / 004.02 region Tenth descriptor vector 168.18 / 199.01 / 026.02 / 177.02 / 172.01 / 005.02 city Eleventh Descriptor Vector 174.19 / 205.01 / 032.03 / 267.03 / 178.01 / 006.02 Street Twelfth Descriptor Vector 180.20 / 211.01 / 132.04 / 033.04 / 184.01 / 007.02 house The third version of the first vector of this 048.21 / 139.03 / 212.01 / 189.01 / 190.01 / 001.01 Subscriber's UIN 192.21 / 151.01 / 188.02 / 189.02 / 190.02 / 004.02 region 198.21 / 157.01 / 194.03 / 195.03 / 196.02 / 005.02 city 204.21 / 163.01 / 200.04 / 201.04 / 202.02 / 006.02 Street 210.21 / 169.01 / 206.05 / 207.05 / 208.02 / 007.02 house Hypercube Phones Fifth descriptor vector 216.22 / 241.01 / 098.02 /117.02 / 220.01 / 008.02 Win phone Sixth descriptor vector 222.23 / 247.01 / 104.04 / 105.04 / 226.01 / 009.02 region telephone code Seventh Descriptor Vector 228.24 / 253.01 / 110.02 / 237.02 / 232.01 / 010.02 area code Eighth Descriptor Vector 234.25 / 259.01 / 104.03 / 105.03 / 238.01 / 011.02 phone number The first version of the second vector of this 240.26 / 205.01 / 260.01 / 243.01 / 244.01 / 008.02 Win phone 246.26 / 211.01 / 242.02 / 243.02 / 244.01 / 009.02 region telephone code 252.26 / 217.01 / 248.03 / 249.03 / 250.01 / 010.02 area code 258.26 / 223.01 / 254.04 / 255.04 / 256.01 / 011.02 phone number Hypercube subscribers Thirteenth Descriptor Vector 264.27 / 277.01 / 032.04 / 033.04 / 268.01 / 008.02 Win phone The fourth version of the first vector of this 048.28 / 175.04 / 278.01 / 273.01 / 274.01 / 001.01 Subscriber's UIN 276.28 / 253.01 / 272.02 / 273.02 / 274.02 / 008.02 Win phone Hypercube phone codes Root vector 282.29 / 091.03 / 314.04 / 315.04 / 286.01 / 000.01 First descriptor vector 288.30 / 319.01 / 368.02 / 351.02 / 292.01 / 012.01 UIN phone code Second descriptor vector 294.31 / 325.01 / 314.04 / 315.04 / 298.01 / 009.01 region telephone code Third descriptor vector 300.32 / 331.01 / 284.03 / 357.03 / 304.01 / 004.01 region Fourth Descriptor Vector 306.33 / 337.01 / 314.04 / 315.04 / 310.01 / 010.01 area code Fifth descriptor vector 312.34 / 343.01 / 362.01 / 291.01 / 316.01 / 005.01 city The first version of the first vector of this 318.35 / 271.01 / 344.01 / 321.01 / 322.01 / 012.01 UIN phone code 324.35 / 277.01 / 320.02 / 321.02 / 322.01 / 009.01 region telephone code 330.35 / 283.01 / 326.03 / 327.03 / 328.01 / 004.01 region 336.35 / 289.01 / 332.02 / 333.03 / 334.01 / 010.01 area code 342.35 / 295.01 / 338.03 / 339.04 / 340.01 / 005.01 city Sixth descriptor vector 348.36 / 379.01 / 314.03 / 315.03 / 352.01 / 012.02 UIN phone code Seventh Descriptor Vector 354.37 / 385.01 / 296.03 / 309.03 / 358.01 / 009.02 region telephone code Eighth Descriptor Vector 360.38 / 391.01 / 302.02 / 357.02 / 364.01 / 004.02 region Ninth Descriptor Vector 366.39 / 397.01 / 314.03 / 315.01 / 370.01 / 010.02 area code Tenth descriptor vector 372.40 / 403.01 / 314.04 / 315.04 / 376.01 / 005.02 city The first version of the second vector of this 378.41 / 331.01 / 403.01 / 381.01 / 382.01 / 012.02 UIN phone code 384.41 / 337.01 / 380.02 / 381.02 / 362.01 / 009.02 region telephone code 390.41 / 343.01 / 386.03 / 387.03 / 388.01 / 004.02 region 396.41 / 349.01 / 392.02 / 393.02 / 394.01 / 010.02 area code 402.41 / 355.01 / 398.03 / 399.04 / 400.01 / 005.02 city

Information retrieval scheme is presented in figure 4.

Information search in the database is carried out in accordance with the search key formed from the following unified values of the measurement coordinates:

measurement "telephone PIN" - the value of the coordinate "**" unknown at the beginning of the search;

Dimension “Last name” - the value of the desired coordinate “Smirnov”;

measurement "City" - the value of the desired coordinate "Petrozavodsk". In this example, the database buffer is conditionally not used. Therefore, the search procedure begins by referring directly to the descriptor vectors recorded in the database file in order to compare the values of the sequence numbers of the Smirnov and Petrozavodsk coordinates.

In accordance with the “Smirnov” coordinate identifier, the subscriber’s hypercube descriptor vector and the offset value of the address pointer of the first type of the identifier of the coordinate identifier of the first version of the given vector are specified in the descriptor vector. They make the transition to this coordinate identifier, using address pointers of the second type, read all coordinate identifiers included in the first version of this vector. Based on the comparison of the sequence numbers of the hierarchy levels, the coordinate identifier is determined with the sequence number of the hierarchy level equal to 1 (unchanged coordinate from the “Subscriber’s PIN” measurement). At the address pointer of the first type of this identifier, the coordinates of the first version of the vector of this transition to the descriptor vector with a matching identifier.

The address pointers of the first type of the descriptor vector and the intermediate versions of the vector of this one carry out a sequential transition to all other versions of the vector of this hypercube of subscribers. The address pointers of the second type read all identifiers of the coordinates of the versions of the vector of the given and compare them with the identifier of the search key by the value of the Petrozavodsk coordinate, which is located in the second modified version of this vector.

In the composition of the third modified version of this vector, the identifier of the measurement coordinate value of the “telephone PIN” is found and the transition to the fifth vector of the phone’s hypercube descriptor is performed on it. Using the address pointer of the first type of the descriptor vector, a transition to the first version of the second vector of this telephone hypercube is performed. Due to the fact that the address pointer of the first type of identifier of the coordinates of the specified version of the given vector has a link to the descriptor vector, the search for versions of data vectors in the phone’s hypercube is completed.

The address pointers of the second type read the identifiers of the coordinates of the first version of the second vector of this telephone hypercube. Coordinate identifiers are replaced by standardized coordinate values from the measurement. The full telephone number of the subscriber is placed in the search report.

INDUSTRIAL APPLICABILITY.

The claimed technical solutions are industrially applicable, because they use well-known industrially produced hardware, well-known industrially produced software tools. The claimed technical solutions can be repeated in the industry using the above description and formula.

Claims (16)

1. A multidimensional database, the logical model of which consists of measurements, each of which corresponds to one of the data types, hypercubes, each of which corresponds to one of a subset of the data, and vectors, each of which corresponds to one of the data, characterized in that
the measurement consists of standardized coordinate values and codes of their identifiers, a hypercube consists of coordinate identifiers connected by address pointers with coordinate identifiers of data vectors with matching codes recorded last in the file;
the vector of this consists of coordinate identifiers associated with address pointers with coordinate identifiers of data vectors with matching codes previously written to the file, while the coordinate identifiers of data vectors written to the file first are connected by address pointers with coordinate identifiers of the hypercube with matching codes,
each of the coordinate identifiers is represented by a set including
the name of the identifier, consisting of the magnitude of its offset relative to the beginning of the file and the serial number of the entries in the vector file;
an identifier code consisting of a digital measurement code and a digital code of a uniform coordinate value in this measurement;
a set of address indicators, and the set of address indicators includes
an address pointer of the first type, consisting of the offset relative to the beginning of the file of the same type of address pointer of the previous coordinate identifier with the same code and the serial number of this coordinate identifier in a cyclic list formed by address pointers of identifiers of coordinates of data vectors and a hypercube;
an address pointer of the second type, consisting of the offset relative to the beginning of the file of the same type of address pointer of the previous coordinate identifier and the sequence number of the coordinate identifier in the cyclic list formed by the address pointers of the coordinate identifiers that are part of one vector of this data, while the first coordinate order in the record is associated with the last coordinate identifier in order of recording;
an address pointer of the third type, consisting of the offset relative to the beginning of the file of the same type of address pointer of the previous coordinate identifier at a higher hierarchy level and the sequence number of the hierarchy level of this coordinate identifier, while the address identifiers of coordinate identifiers that are part of the same hierarchical groups of a given vector form a finite list;
an address pointer of the fourth type, consisting of the offset relative to the beginning of the file of the same type address pointer of the coordinate identifier included in one variant group, or, in the case of the completion of one variant group, which is part of another variant group, and the serial number of the variant group, with address pointers identifiers of coordinates, which are part of one vector of this, form a finite list. 2. The multidimensional database according to claim 1, characterized in that each hypercube coordinate identifier is represented as a descriptor vector, while the address pointer of the first type of the descriptor vector coordinate identifier contains an offset value relative to the beginning of the file and the serial number of the address pointer of the first type of vector coordinate identifier given with the matching identifier code, written to the file last. 3. The multidimensional database according to claim 2, characterized in that;
descriptor vectors in the hypercube are logically combined into two or more segments;
a root vector consisting of one coordinate identifier with a system code is included in the composition of each segment of the hypercube;
the root vectors of the hypercubes are connected into a cyclic list by address pointers of identifiers of coordinates of the first type, the serial numbers of which take the values of the numbers of the hypercubes in the order they are written to the file;
the root vector and descriptor vectors that are part of one hypercube segment are connected to the index structure of a binary tree type by address pointers of coordinate identifiers of the second and third types;
root vectors and descriptor vectors with matching coordinate identifier codes included in different segments of the same hypercube are connected into cyclic lists by address identifiers of coordinates identifiers of the fourth type, while the sequence numbers of the address identifiers of coordinate identifiers take the values of the numbers of segments of the hypercube in the order they are written to the file. 4. The multidimensional database according to claim 3, characterized in that it includes a buffer, for example, located in the computer’s RAM, consisting of tables of values of descriptor vectors with the largest ordinal numbers of address pointers of coordinates of the fourth type and a list of tables of values of vectors descriptors in the buffer, while
the list of tables of descriptor vector values contains the addresses of the tables in the buffer and is formed in the order in which the tables are written to the file;
each table of descriptor vector values corresponds to one of the dimensions;
rows of tables of descriptor vector values correspond to the coordinates of measurements, are formed in the order of writing standardized coordinate values to a file and contain sets of fields, each of which corresponds to one of the hypercubes formed by measurements corresponding to the tables;
one set of field fields for descriptor vector value table rows includes a field for the offset value of the name of the descriptor vector identifier relative to the beginning of the file, an ordinal number field of the address pointer of the fourth type of identifier for the vector of the descriptor, an offset value field for the start of the file of the address pointer of the first type of identifier of the first type of the vector coordinate identifier of the given vector with the same the last code written to the file, and the field of the serial number of the address pointer of the first type of coordinate identifier nats of this vector given. 5. The multidimensional database according to claim 4, characterized in that the buffer contains the vectors of data involved in the operations of reading and writing information, and inverted lists of coordinates of these data vectors,
inverted lists of coordinates of data vectors contain the addresses of the data vectors in the buffer;
each vector of a given has a serial number equal to the largest serial number of the record in the file of one of the identifiers of the coordinates of the vector of this. 6. The multidimensional database according to claim 5, characterized in that the rows of the tables of descriptor vector values additionally contain fields of unified coordinate values, index fields of unified coordinate values and address fields of inverted coordinate lists of data vectors in the buffer. 7. The multidimensional database according to claim 1, characterized in that the vector of this data may contain one or more fragments of unstructured information, each of which is limited on two sides by coordinate identifiers that determine the data type and recording format of a fragment of unstructured information. 8. The method of managing the multidimensional database according to claim 1, characterized by the presence of operations of identification, writing, reading, changing, deleting and compressing information, controlling the internal consistency of information during the joint work of two or more users, restoring the internal consistency of information after an emergency shutdown,
characterized in that
information identification is provided by digitally encoding standardized coordinate values as part of measurements and replacing actual coordinate values by digital codes as part of data vectors;
writing, reading, changing, deleting and compressing information is provided through the formation of logical sections of multidimensional space, represented by cyclic lists of identifiers of coordinates of data vectors and hypercubes with matching identifier codes;
control of the internal consistency of information during the joint work of two or more users is ensured by sequential numbering of data vectors in the order they are written to a file;
the restoration of internal information consistency after the multidimensional database crashes is ensured by means of leading recording of coordinate identifiers in the composition of descriptor vectors relative to the moment of their recording in the composition of data vectors. 9. The method of managing a multidimensional database according to claim 8, characterized in that the identification of information is performed in the following sequence:
at the first step, the actual values of the coordinates of the data received from an external source of information are compared with the standardized values of the coordinates as part of the measurements;
at the second step, in case of discrepancy between the actual coordinate values of the given and the standardized coordinate values in the measurements, the actual values are included in the measurements as new standardized values with the assignment of digital identifier codes to them;
in the third step, the actual coordinate values of the given are replaced with digital codes of identifiers of standardized coordinate values as part of the measurements. 10. The method of managing a multidimensional database according to claim 8, characterized in that the compression of information is carried out in the following sequence:
at the first step, a second instance of the hypercube is created, used to record newly received information and associated with the first instance by address pointers of the fourth type of root vectors and descriptor vectors;
at the second step, a copy of the first instance of the hypercube is created that is not associated with the first instance with address pointers of the fourth type of root vectors and descriptor vectors, used as a receiver of information previously recorded in the first instance of the hypercube - the source of information;
in the third step, information is compressed by generating consolidated versions of the data vectors contained in the information source and transferring them to the information receiver;
in the fourth step, the information source is replaced by the information receiver by linking the second instance and a copy of the first instance of the hypercube with address pointers of the fourth type of root vectors and descriptor vectors, the first instance of the hypercube is deleted from the file. 11. The method of managing a multidimensional database according to claim 9, characterized in that the restoration of internal information consistency after an emergency shutdown is performed in the following sequence:
at the first step, based on the comparison of the values of the vectors of the descriptors of the hypercube, they move to the identifier of the coordinate of the last vector of the given one, the most offset from the beginning of the file, after that, using the address pointers of the second type, go to other identifiers of coordinates of the last vector of this;
moreover, if the last vector of this complete set of coordinate identifiers is present, the restoration of internal information consistency after an abnormal termination of work is completed, if the last vector does not contain at least one of the coordinate identifiers, the location address of which is recorded in the descriptor vector, additionally next steps;
at the second step, the values of the descriptor vectors, the identifier codes of which correspond to the identifiers of the coordinates of the previous data vectors, are replaced by the offset values relative to the beginning of the file and the sequence numbers of the address pointers of the first type of identifiers of the coordinates of the vector of this;
at the third step, the transition to the last identifier of the coordinate of the penultimate vector of this offset from the first identifier of the coordinate of the last vector of this data by the value specified by the format of the coordinate identifier is carried out, then the address identifiers of the second type go to the remaining identifiers of coordinates that are part of the penultimate vector of this;
in the fourth step, the codes of the coordinates of the penultimate vector of the given vector are compared with the codes of the coordinates of the identifiers of the descriptor vectors corresponding to the codes of the coordinate identifiers that are not present in the last vector of the given, and if the codes coincide, the values of the vectors of the descriptors are replaced by the offset values and serial numbers of the address pointers of the first type of identifiers of the penultimate coordinates vector of this, when the codes do not match, go to the next in order of approximation to the beginning of the file to the data vectors up to finding the identifiers of coordinates whose codes coincide with the codes of the identifiers of coordinates that are not present in the last vector of this data, with the subsequent replacement of the values of the descriptor vectors;
in the fifth step, the last vector of the given one with an incomplete set of coordinate identifiers is deleted from the file. 12. The method of managing a multidimensional database according to claim 8, characterized in that the reading of information is performed in the following sequence:
in the first step
the user generates a search key from one or more coordinate identifiers;
as a part of the buffer, they search for descriptor vector values with the corresponding coordinate identifiers by finding descriptor vector tables in the address list whose positions correspond to digital measurement codes, go to the indicated tables and search for strings in their structure whose positions correspond to digital codes of standardized coordinate values;
if there is no value in the tables of at least one of the descriptor vectors with the coordinate identifiers corresponding to the search key, the reading is stopped, otherwise the descriptor vector values are read;
comparing the serial numbers of address pointers of the fourth type of identifiers of coordinates of the descriptor vectors;
if they are equal to the sequence number corresponding to the last segment of the hypercube, the values of the found descriptor vectors are compared with each other by the value of the sequence numbers of the address pointers of the first type, followed by the choice of the descriptor vector with the lowest number, otherwise, the address pointer of the fourth type proceeds to the next segment of the hypercube ;
in the second step
as part of a hypercube segment, they search for other descriptor vectors whose coordinate identifiers coincide with the search key by traversing the binary tree of the index structure;
if the sequence numbers of address indicators of the fourth type are equal, the descriptor vectors are compared with each other by the value of the sequence numbers of address indicators of the first type with the subsequent selection of the descriptor vector with the lowest number; otherwise, the address pointer of the fourth type moves to the next segment of the hypercube until the sequence numbers match address pointers of the fourth type of descriptor vectors;
in the third step
based on the value of the address pointer of the first type of the descriptor vector with the lowest sequence number, the identifier of the coordinate of the vector of the given is read;
in accordance with the values of the address pointers of the second type, the remaining identifiers of the coordinates of the vector of this data are read and compared with the identifiers of the coordinates of the search key;
if the identifiers of the coordinates of the vector of the given coincide with the identifiers of the coordinates of the search key, one of the matching identifiers of the coordinates of the vector of the given with the highest serial number of the address pointer of the second type is determined;
otherwise, they proceed to the fifth step of the current sequence of reading information;
in the fourth step
bypass hierarchical and variant groups of a given vector, starting with the matching coordinate identifier with the highest sequence number of the address pointer of the second type, if the coordinates of the search key are included in the same hierarchical and variant groups of this vector, it is included in the search report;
in the fifth step
as a part of the read vector, a comparison is made between each other in terms of the sequence numbers of the address pointers of the first type of coordinate identifiers corresponding to the search key, the selection of the coordinate identifier with the smallest largest serial number, reading of the coordinate identifier of the previous vector given by the address pointer of the first type and reading of the remaining coordinate identifiers previous vector given by address pointers of the second type;
moreover, the fourth and fifth steps are repeated in order of approximation to the beginning of the file until reading one of the identifiers of the coordinates of the vector of the data corresponding to the search key, with the serial number of the address pointer of the first type equal to one. 13. The method of managing a multidimensional database according to claim 12, characterized in that the control of the internal consistency of information during the joint work of two or more users with the same data is carried out in the following sequence:
at the first step, users independently read information, which begins by reading in the buffer a consolidated version of a given vector corresponding to a search key; if there is no consolidated version of a vector in a buffer in the buffer, reading is done from a file;
at the second step, users independently create modified (deleted) versions of the given vector, which are temporarily assigned the record number of the previously read consolidated version of the given vector;
at the third step, by a command to record the user who first completed the formation of the changed (deleted) version of the vector of the data, they read again in the buffer or file of the consolidated version of the vector of this data, while the execution of the commands for writing the remaining users is blocked;
in the fourth step, the sequence numbers of the records of the changed (deleted) and re-read consolidated versions of this vector are compared among themselves;
moreover, if the sequence numbers of the records coincide, the modified (deleted) version of the vector of the data is written to the file, and if the serial numbers of the records do not match, the changed (deleted) and new consolidated versions of the vector of the data are re-formed with the current serial number of the record, the new consolidated version replaces in the buffer the previous consolidated version, the modified (deleted) version is written to a file, the execution of write commands is unlocked with the repetition of the third and fourth steps for each next user;
in the fifth step, the last consolidated version of this vector is removed from the buffer if there are no read or write commands for a given period of time. 14. The method of managing the multidimensional database of claim 8, characterized in that the information is recorded in the following sequence:
at the first step, this is converted into a vector form by generating coordinate identifiers, linking them with address pointers of the second, third and fourth types, determining by calculation the values of address pointers of the first type if there are no descriptor vectors with matching identifier codes in the hypercube, or by copying the address values pointers of the first type of descriptor vectors if there are descriptor vectors with matching identifier codes in the hypercube ators;
at the second step, if there are no descriptor vectors with matching identifiers in the hypercube, write these descriptor vectors to the file, the address pointers of the first type of which have the calculated values of the same type of address pointers of the given vector, or if the descriptor vectors in the coordinate block contain the change to calculated values of the offset values relative to the beginning of the file and the serial number of the record of the same type of address pointers of descriptor vectors;
in the third step, the vector of the data is written to the end of the file. 15. The method of managing a multidimensional database according to claim 14, characterized in that the information is changed at a logical level by writing several versions of the vector given in the following sequence:
at the first step, the current version of the vector of the given is read;
at the second step, an altered version of the vector of the given one is composed of one or more identifiers of coordinates and / or fragments of unstructured information changed relative to the current version, and one invariable identifier of the coordinate;
at the third step, a modified version of the vector of the data is recorded with the coordinates of the offset values relative to the beginning of the file recorded in the first version of the vector of the data stored in the names of the identifiers;
moreover, the subsequent reading of the changed information is carried out by forming a consolidated version of this vector, consisting of one coordinate identifier with a constant identifier code and one or more last-in-order records of the changed coordinate identifiers and / or fragments of unstructured information. 16. The method of managing a multidimensional database according to clause 15, characterized in that the information is deleted at a logical level by writing several versions of the vector given in the following sequence:
at the first step, the current version of the vector of the given is read;
at the second step, a remote version of the vector of the given is formed as a part of one or several identifiers of coordinates removed from the current version of the vector of the given and one constant identifier of the coordinate;
at the third step, a remote version of the vector is recorded, replaced with zero values of the codes of the remote coordinate identifiers;
moreover, the subsequent reading of the stored information is carried out by forming a consolidated version of this vector, consisting of coordinate identifiers with nonzero values of the coordinate identifier codes and fragments of unstructured information enclosed between them.

Images