Glossary

A privilege required to access schema objects (base tables and viewed tables). Before you can use SQL statements to search the data in a schema object or use commands to manipulate a schema object, you must have access privileges for that schema object. There are several types of access privilege, including the SELECT privilege, INSERT privilege, and IMPORT TABLE privilege.

A base table for which the background-import facility and chunk archiving function can be used.

A directory that stores archive files.

A file that stores archived data.

A column that is used to narrow down the search range when an archivable multi-chunk table is searched. An archive range column must be defined for each archivable multi-chunk table. A column that includes datetime data can be used as an archive range column.

The state in which archived data is stored in an archive file.

Audit privilege is a collective term for the following two privileges. You must have an audit privilege to use the audit trail facility.

• Audit admin privilege

  The privilege that an HADB user must have to perform operation of the audit trail facility.

• Audit viewer privilege

  The privilege that an HADB user must have to reference audit trails.

An audit target definition defines the combinations of events, HADB users, and schema objects for which to output audit trails. The information defined by a CREATE AUDIT statement is an audit target definition.

An operation for which an audit trail is output. There are two types of audit target events, mandatory audit events and optional audit events. For mandatory audit events, an audit trail is always output if the audit trail facility is enabled. For optional audit events, auditors can choose whether to output audit trails.

A facility that outputs information about activity by HADB users, such as database access and command execution, to a file as audit trail data. For example, when HADB users access a table, information about the operations they perform is output as an operation record (audit trail). This might include the time at which the users accessed the table, their authorization identifiers, the operations they performed, and the schema object on which they performed the operations. By viewing the output audit trail, an administrator can find out information such as who accessed what schema object at what time, and what operations they performed.

A file to which audit trail data is output that records information about activity by HADB users, such as database access and command execution.

An HADB user who has an audit privilege. An auditor is responsible for such tasks as operating the audit trail facility and using the output audit trails to audit database usage.

A user ID that is allowed to connect to HADB.

An optional facility of the adbimport command. Applying this facility when you execute the adbimport command enables you to concurrently execute a data search in a table and import data into the same table.

If a branch row was created, the row that contains the information about the branching destination. If no branch row was created, one row of data is stored in a base row.

Variable-length column data (row) stored on a page other than the page on which the base row is stored.

A function that allows an administrator to centrally manage the client definition of each HADB client from the HADB server.

A unit of data that is stored in a base table by a single background-import operation. A chunk consists of base table data that was stored by background import and the index data that was created during background import. A new chunk is created each time background import is executed.

A chunk can be in any of the following statuses:

• Normal status

• Wait status

• Delete-pending status

Note that multiple chunks can be merged into a single chunk (chunk merging).

To compress the data in a chunk by using the chunk archiving function and store the compressed data in an archive file.

A function that compresses the data in a chunk and stores the compressed data in an archive file.

An identifying number assigned to each chunk. Each chunk ID number is unique within the base table.

When a chunk is in normal status, the data contained in it can be manipulated by a data manipulation SQL statement. A chunk in normal status is created when background import is executed. You can use the adbchgchunkstatus command to change the status of a chunk from normal status to wait status.

When a chunk is in wait status, the data contained in it cannot be manipulated by a data manipulation SQL statement. However, a chunk in wait status can be deleted by using the PURGE CHUNK or TRUNCATE TABLE statement. A chunk in wait status is created when you specify the status of the chunk to be created while performing background import. You can use the adbchgchunkstatus command to change the status of a chunk from wait status to normal status. Note that a chunk in wait status does not become the current chunk.

The operation of merging (combining) multiple chunks into a single chunk. The adbmergechunk command is used to merge chunks.

Multiple chunks designated to be merged are called merge-source chunks, and the single chunk into which multiple chunks are merged is called the merge-target chunk.

To decompress the data in an archive file by using the chunk archiving function and store the decompressed data in the data DB area. At this time, an index is also re-created.

A definition that specifies the execution environment of the HADB client.

A directory that stores a group of files related to a client process.

A facility for grouping multiple HADB clients together and specifying ranges for the numbers of connections and processing real threads that can be used by the group.

A system configuration in which a standby server machine (standby system) is provided in addition to the server machine (active system) that performs application processing. If the active system cannot continue application processing because of a failure, the standby system takes over as the active system and continues the application processing.

A format in which data is stored in the database at the column level. In column store format, the data in each column of a table is stored together in the database.

A table defined with column store format selected as the table-data storage format. There are two table-data storage formats: row store format and column store format. You can specify the table-data storage format when defining the table.

A file used to manage the status of commands, such as executing status and interrupted status.

Audit trail files in a format in which the information can be collected, viewed, and managed with JP1/Audit. When audit trails that have been output to audit trail files are converted by using the adbconvertaudittrailfile command, the converted audit trails are output to common format audit trail files.

A user privilege needed by an HADB user in order to connect to the HADB server. This is a prerequisite user privilege for using the HADB server. Without the CONNECT privilege, a user cannot connect to the HADB server even if he or she has other privileges.

An object that manages connections (between HADB clients and the HADB server).

A function that performs a text data search, ignoring the differences between uppercase and lowercase letters, between half-width and full-width characters, and between Japanese hiragana and katakana characters.

Information related to retrieval methods that is collected by the HADB server from base tables, B-tree indexes, and text indexes for the purpose of optimizing table retrieval processing. Executing the adbgetcst command collects cost information and stores it in a system table.

Among the created chunks, this is the chunk to which a user is adding data. Only one current chunk exists in a base table.

The current chunk is changed in the following cases:

• When a new chunk is created using background import

• When multiple chunks are merged, including the current chunk

The current chunk might also be changed in the following case:

• When the status of a chunk is changed

A DB area that stores tables and indexes.

A process of storing a full table's worth of data in one step. The adbimport command is used to import data.

An SQL statement used to manipulate table data (retrieval SQL statements and update SQL statements).

The process of exporting table data and then importing the exported data, for the purpose of deleting the invalid row data in the table. Data reorganization can be performed for an entire table (table reorganization) or for each chunk (chunk reorganization).

A logical area that stores tables, indexes, and their definition information. DB areas are classified into the following according to their uses:

• Master directory DB area

• Dictionary DB area

• System-table DB area

• Work table DB area

• Data DB area

A file that makes up a DB area.

A directory that stores a group of database-related files, such as DB area files and system log files.

A user privilege required for managing HADB users and privileges (user privileges and the schema operation privilege).

A generic term for SQL statements used for the following purpose:

• Defining and deleting schemas, tables, and indexes

• Creating and deleting HADB users

• Granting privileges to HADB users

• Revoking HADB users' privileges

The SQL statements described under Definition SQL in the manual HADB SQL Reference qualify as definition SQL statements.

An unnecessary chunk (merge-source chunk) that remains undeleted because a chunk-merging process was interrupted. An unnecessary chunk (data before reorganization) that remains undeleted because reorganization of a system table (base table) was interrupted also applies.

The data contained in this chunk cannot be manipulated by a data manipulation SQL statement. However, this chunk can be deleted by using the PURGE CHUNK or TRUNCATE TABLE statement. Note that the status of this chunk cannot be changed by using the adbchgchunkstatus command.

An access privilege granted to an HADB user by another HADB user.

Example:

• The access privileges HADB user A has for table X.T1 shall be called P1.

• The access privileges HADB user A grants to HADB user B for table X.T1 shall be called P2.

• The access privileges HADB user B grants to HADB user C for table X.T1 shall be called P3.

In this scenario, the dependent privileges of the access privileges P1 are the access privileges P2 and P3. The dependent privileges of the access privileges P2 are the access privileges P3. The access privileges P3 do not have any dependent privileges.

For example, suppose that the access privileges P1 are revoked. In this case, access privileges P2 and P3 which are dependent privileges are also revoked. If the access privileges P2 are revoked, the access privileges P3 which are dependent privileges are also revoked.

A viewed table that is affected when, for example, a base table is deleted by using the DROP TABLE statement or when a viewed table is deleted by using the DROP VIEW statement. The following figure shows an example of dependent viewed tables.

Explanation

• Viewed tables V1, V2, and V4 are dependent on base table T1.

• Viewed tables V1, V2, and V4 are dependent on base table T2.

• Viewed tables V2 and V4 are dependent on base table T3.

• Viewed tables V3 and V4 are dependent on base table T4.

• Viewed tables V2 and V4 are dependent on viewed table V1.

• Viewed table V4 is dependent on viewed table V2.

• Viewed table V4 is dependent on viewed table V3.

• There is no dependent viewed table for viewed table V4.

Operations other than deletion of a base table or viewed table might also affect dependent viewed tables. For details about the operations that affect dependent viewed tables, see (3) Invalidating viewed tables in 2.1.2 Viewed tables.

A DB area that stores dictionary tables and their indexes.

A viewed table that is created from a dictionary table (base table) for search purposes. The schema name of a dictionary table is MASTER.

A table that stores database definition information, such as table definitions, column definitions, and index definitions. An HADB user cannot search a dictionary table (base table). To check the information stored in a dictionary table (base table), the user must search the dictionary table.

The directory in which synonym dictionary files are stored. The directory for storing synonym dictionary files is the directory specified in the adb_syndict_storage_path operand in the server definition.

A column that uses a search condition in a subquery to reference a table specified in the FROM clause of an external query. Referencing a table by using a search condition in a subquery is called external reference. The following figure shows an example of an external reference column.

A column (or a combination of multiple columns) that references the primary key of another base table. Primary keys and foreign keys can be used to define relationships among tables in a database.

A buffer used for inputting/outputting data stored on a disk in DB areas. Global buffers are allocated in shared memory.

A privilege that allows an HADB user to grant access privileges to another HADB user.

An OS user who manages HADB. The HADB administrator is the owner of the server directory and the DB directory, and can execute all HADB commands.

An HADB-specific group that is set in the server machine's OS. The HADB administrators group becomes the primary group to which the following OS users belong:

• HADB administrator

• OS user who belongs to the HADB administrators group

See OS user who belongs to the HADB administrators group.

A client-side system in HADB that operates as part of the client-server system.

A server-side system in HADB that operates as part of the client-server system.

A user who, by connecting to the HADB server, can take actions such as importing data, retrieving or updating data, and managing HADB users. Privileges needed to execute various types of operations can be granted to an HADB user.

Processing that uses a hash table area. The following types of processing apply:

• Hash join as a table joining method

• Global hash grouping as a grouping method

• Hash execution as a method for processing subqueries

• Hash execution as a method for processing SELECT DISTINCT

• Hash execution as a method for processing the set operation

A function of HA Monitor that protects the data on a shared disk. You must select one of the following as the shared disk data protection method used by HA Monitor. We recommend that you select host reset.

• Host reset

• SCSI reservation for shared disk

For details about host reset and SCSI reservation for shared disk, see Host reset and SCSI reservation for shared disk in the manual HA Monitor for Linux(R) (x86).

The operation of rebuilding an index defined for a base table. The adbidxrebuild command is used to rebuild an index.

A directory that is created when the HADB server and HADB clients are installed.

The data from a deleted row or the original data for a row that has since been updated is not deleted from the disk. This data that is present on the disk but is not subject to retrieval is called invalid row data. If you repeatedly perform operations that delete and update rows, the amount of invalid row data in the table increases. If invalid row data increases, it might affect as follows:

• Data storage efficiency drops due to the increase in the amount of invalid row data.

• Retrieval performance slows down because the number of pages to be referenced increases during retrieval processing.

A table that is created automatically by the HADB server when the data in chunks is archived. A location table stores the following types of information:

• Chunk ID

• Archive file path

• Range of values (maximum and minimum values) in the archive range column of each archive file

Location tables are used by the HADB server and are not used by users.

One location table is created for each archivable multi-chunk table.

Internal system information used by HADB for managing the various types of information related to DB areas.

A DB area that stores the master directory.

The node containing the HADB server that controls the multi-node function. The master node corresponds to the active system in HA Monitor.

The HADB server on the master node accepts application and command connection requests and distributes processing to slave nodes with lighter loads.

Update SQL statements and commands such as the adbimport command are processed on the master node.

Action taken when a node failure occurs on the master node so that the master node is separated from the multi-node configuration and one of the slave nodes becomes the master node.

A message object file that stores the test of the messages that are output by HADB.

A file that stores the messages that are output by HADB.

A base table in which multiple chunks can be created. For a multi-chunk table, the background-import facility can be used. Multi-chunk tables can be categorized into two types: regular multi-chunk tables and archivable multi-chunk tables.

A system configuration consisting of multiple HADB servers that are using the multi-node function.

This function realizes the load distribution of retrieval SQL statement processing by coordinating multiple HADB servers.

A directory that stores the synonym dictionary files used to perform synonym searches when using the multi-node function. The multi-node synonym dictionary storage directory is the directory specified in the adb_syndict_node_storage_path operand in the server definition.

A node is a server machine on which an HADB server is installed, including the operating system, HADB server, and HA Monitor installed on that server machine. A node corresponds to a host in HA Monitor.

In this manual, a node failure means a failure that requires a node to be separated from a multi-node configuration. A node failure corresponds to a host failure or a server failure in HA Monitor.

A slave node in which a node failure occurred is disconnected from the multi-node configuration. This is called node separation.

A constraint that does not permit null values for column values.

A user who logs on to the OS and can utilize the OS functions. In HADB, the following three types of OS users operate the HADB server:

• Superuser

• HADB administrator

• OS user who belongs to the HADB administrators group

An OS user who manages HADB separately from the HADB administrator. The OS user can access the server directory and DB directories owned by the HADB administrator. The OS user can also execute some of the HADB commands.

A database search method that is used by HADB. In out-of-order execution, a search process is divided into row units and executed in parallel.

One of the units used for storing data. It is the smallest unit for disk I/O operations. When data is read from or written into a disk, it is done in page units.

A column (or combination of columns) that uniquely identifies rows in a table. Defining a primary key allows you to maintain the uniqueness of the data in a base table. When you define a primary key, the following constraints apply to the columns comprising the primary key:

• Uniqueness constraint

• NOT NULL constraint

An HADB user needs privileges in order to perform various types of operations on an HADB server. HADB privileges can be roughly classified into the following three types:

• User privileges

• Schema operation privilege

• Access privilege

An HADB user can have multiple privileges. An HADB user who does not have the privileges corresponding to a certain operation cannot execute that operation.

A real thread that is allocated when the HADB server starts, and that is used for executing SQL statements (data import, for example).

A thread that is created by HADB in order to execute, in parallel, multiple processes in a single real thread. Processes are executed in parallel by executing multiple pseudo threads in a single real thread (only one pseudo thread can actually be executing at any time).

An index that manages the minimum and maximum values (a range) in the column of data stored in chunks and segments of a table.

A thread managed by the OS (kernel and library).

A base table for which a primary key is defined, and which is referenced from a referencing table using a foreign key.

A base table that defines a referential constraint and a foreign key.

A constraint that maintains a relationship between tables (referential consistency of data). You can define a referential constraint for a specific column (foreign key column) when you define a base table.

A base table for which the background-import facility, which is a basic function for multi-chunk tables, can be used. For regular multi-chunk tables, the chunk archiving function cannot be used.

The process of restoring the data in unload files to a system table as part of the reorganization of a system table.

An SQL statement used to search table data. The SELECT statement qualifies as a retrieval SQL statement.

Action taken to return a node that is separated from the multi-node configuration (due to a node failure) to the multi-node configuration (after correction of the failure). Nodes can be returned to the multi-node configuration as only slave nodes.

A format in which data is stored in the database at the row level. In row store format, one row of data is stored in the database as one record.

A table defined with row store format selected as the table-data storage format. There are two table-data storage formats: row store format and column store format. You can specify the table-data storage format when defining the table.

A logical concept that includes tables and indexes. In HADB, each HADB user can own only one schema.

A schema operation privilege that an HADB user needs in order to define or delete schemas, tables, and indexes.

An element that can be defined in a schema. A table or index is an example of schema object.

A privilege needed for managing an HADB user's schemas. The following is the only type of schema operation privilege:

• Schema definition privilege

A function of HA Monitor that protects the data on a shared disk. You must select one of the following as the shared disk data protection method used by HA Monitor. We recommend that you select host reset.

• Host reset

• SCSI reservation for shared disk

For details about host reset and SCSI reservation for shared disk, see Host reset and SCSI reservation for shared disk in the manual HA Monitor for Linux(R) (x86).

A unit for storing data that is used when allocating areas in files to tables and indexes. A segment consists of contiguous pages.

Segment size means the number of pages in a segment.

A directory that stores the files needed by the HADB server for execution. It stores a group of files related to the server process.

A base table in which only one chunk can be created. For a single-chunk table, the background-import facility and chunk archiving function cannot be used.

A node that serves as the location of an HADB server that uses the multi-node function to process retrieval SQL statements and some commands. A slave node corresponds to HA Monitor's standby system.

If the master node fails, one of the slave nodes becomes the master node.

A set of character codes that are used to sort or compare characters as defined in the ISO/IEC 14651:2011 standard. Correction search uses this as the basis of notation correction.

A processing real thread that is used when executing SQL statements.

Historical information about the execution of SQL statements that is collected by SQL tracing. The SQL trace information includes the execution time of SQL statements and access path information. SQL trace information is used for the following purposes:

• Determining the causes of errors in SQL statements

• Tuning SQL statements

A function for outputting historical information about the execution of SQL statements (SQL trace information) to files (SQL trace files).

An object for managing the SQL statements to be executed.

A file to which part of the HADB server's operation information is output. The output information can be checked by using the adbstat command. The following type of information is output:

• SQL statement statistical information

A file for managing the operational statuses and termination modes of the HADB server.

A dictionary in which synonyms are listed.

A file storing the information about synonym dictionaries. The HADB server looks in the synonym dictionary file when performing synonym search.

A group of synonyms that are targeted by synonym search.

A file that contains a list of synonyms to be registered in a synonym dictionary. When synonym search is used to search for a word, all the synonyms registered as the synonym group of that word in the synonym list definition file are also used as the search strings.

A function that you can use when searching literatures, research papers, and other document data for a specific word. This function allows you to search also for synonyms of the word at the same time.

Database update history information that is automatically acquired by HADB. The system log is used for recovering a database after an error has occurred in it.

A file to which system logs are output.

A viewed table that is created from a system table (base table) for search purposes. The schema name of a system table is MASTER.

A table for storing the following types of information:

• Cost information for base tables, B-tree indexes, and text indexes

• Chunk information for multi-chunk tables

An HADB user cannot search a system table (base table). To check the information stored in a system table (base table), the user must search the system table.

The process of releasing areas of invalid row data in system tables.

The DB area for storing system tables and indexes of system tables.

A method of retrieving base tables without using B-tree indexes or text indexes.

Data in tabular format that has been derived by the following system-defined functions:

• ADB_AUDITREAD function

• ADB_CSVREAD function

An index that manages information about the position of data (character strings) stored in tables. A text index can be used when character string data to be retrieved contains a specified character string.

A table that becomes the base for a viewed table. The table specified in the query expression of a CREATE VIEW statement is the underlying table.

A constraint that does not permit duplicate column values (or combinations of column values from multiple columns).

A process of outputting, to an unload file, system table data from which invalid row data has been excluded.

A temporary file created during reorganization of a system table.

An SQL statement used to add data to a table, deleting data from a table, or updating data in a table. The following SQL statements qualify as update SQL statements:

• INSERT statement

• DELETE statement

• UPDATE statement

• PURGE CHUNK statement

• TRUNCATE TABLE statement

This is an HADB facility that automatically converts data stored in row store format in a column store table to column store format. When the INSERT or UPDATE statement is executed for a column store table, the added or updated data is stored in row store format in the column store table. If this facility is enabled, the data stored in row store format is automatically converted to column store format.

A privilege that an HADB user needs in order to manage HADB users or connect to the HADB server. The following user privileges are available:

• DBA privilege

• CONNECT privilege

A value that indicates the hierarchical depth of a defined viewed table from the base table. The following figure shows an example of view levels.

Explanation

• The view level of viewed table V1 is 1 because only base tables underlie V1.

• The view level of viewed table V2 is 2 because a base table and a viewed table at view level 1 underlie V2.

• The view level of the viewed table V3 is 1 because only a base table underlies V3.

• The view level of viewed table V4 is 3 because the underlying tables of V4 are at view levels 1 and 2.

A virtual table created by defining the result of the query expression specified in the CREATE VIEW statement as a new table. For example, you can define the result of a set operation for multiple tables as a viewed table, or can define the specific rows or columns in a table as a viewed table.

A function that quickly searches English-language text data for specific English words specified as search keys. There are two methods of word-context search as follows:

• Complete-match retrieval

• Leading-match search

A DB area that stores the tables (work tables) HADB creates internally when executing SQL statements.