back out feature The ability to undo an ODXFIND keyword search and to restore the previous internal ID list.

bit Binary digit. Characters “1” and “0” A measure of computer recording capacity A computer term for referring to the machine instruction of one or zero, setting an electrical switch to on or off.

bitmap index An indexing option which stores index pointer as a bitmap, providing efficient indexing for “low cardinality” data. See also: Cardinality.

bitmap join Bitmap joins are characterized by having the “from” table and the “to” table as the same child, the purpose being to change which owner is current for the table. This step may be taken when joining between a parent and a child that are not already pre-joined.

block A unit of data or computer memory written to a computer disk or other digital storage media.

Boolean operators AND (,), OR (+), and NOT (,-) operators used to combine keywords during an Omnidex retrieval.

B-tree A type of index optimized for sorted and partial key retrievals by on column or concatenation of columns, requiring left justification of data and case-sensitive matching. The fundamental type of indexing on which relational databases are founded.

calling errors Syntax errors that cause programs to fail.

cardinality A term indicating the number of: unique values in a column rows in a table

With respect to databases, there are two distinct types of cardinality: table cardinality and column cardinality.

Table cardinality is the number of rows in a table. Column cardinality is the number of unique values in a column. Omnidex uses these values to determine the best method to use to process a query against multiple tables.

For example:

The products table is a parent table with a primary key that has a column cardinality of 10,000. This means it has 10,000 unique product codes.

The orders table is a child table with 10,000,000 rows. This means it has a table cardinality of 10,000,000.

SELECT O.PRODUCT_CODE, O.CUSTOMER_NO, O.STATUS FROM PRODUCTS P, ORDERS O WHERE P.PRODUCT_CODE=O.PRODUCT_CODE AND P.PRODUCT_NAME = 'PRINTER'

cartesian product A dataset consisting of every possible combination of rows from mulitple tables. For example: Customers is the parent table and activity and orders are the child tables. Selecting from both the activity and orders table for a particular customer will produce a cartesian product. Meaning, every row in the activity table will be qualified with every row in the orders table. If activity has 10 rows for one customer and orders has 10 rows for the same customer, the cartesian product is 100 rows. table1 (t1) contains 2 rows (r1, r2) and table2 (t2) contains 3 rows (r1, r2, r3). The cartesian product or these tables contains 6 rows: t1r1-t2r1, t1r1-t2r1, t1r1-t2r3, t1r2-t2r1, t1r2-t2r2, t1r2-t2r3.

chained list A list containing data elements where each data element has a pointer to the previous and successive data element. See also: linked list

child A child table or detail data set.

child key A column or field in a child table that identifies rows in the table containing the same values; also known as a repeating key. See also: key

child table A table subordinate to a parent table or master data set. A table where several rows may share the same identifying value for any given column.

Child rows are not usually uniquely identifiable by the contents of a single field. There is a one-to-many relationship between the parent table and the child table.

client The user interface aspect of a client/server system, usually a personal computer or small computer workstation dedicated to immediate, local user interactions, such as system navigation and help, rather than widely accessed or centralized computational activity, such as data processing.

client/server A distributed technology approach, or system, where computer processing is divided by function to take advantage of processing on multiple computer units, dividing tasks between client activities and server activities. See client and server.

codified data Data represented by codes instead of actual data values, used by data warehouses to conserve data storage space, e.g., income ranges represented by a single character.

column A logical representation of a field in individual rows for a given table or view. A column defines a set of fields that share the same data and the same position (offset and length) for any given row in a table. column cardinality The number of unique values in a column. See also: cardinality, table cardinality composite index A type of Omnidex index that lets you index data from either a part or a combination of parts of columns, or entire columns (one or more items from a field). You can create composite indexes in DBINSTAL during Omnidex installation. See also: composite keys composite keys A logical MDK or ASK field comprised of several fields or parts of fields. See also: composite index connection A direct link between the application and an environment catalog. container application A computer program, typically ODBC-compliant, that processes data extracted from a database. Generic applications such as Microsoft Excel, Seagate's Crystal Reports and Brio's BrioQuery, that make generic calls through ODBC to a database. Applications that work with a database through an ODBC driver.

criteria count A count of the number of rows qualified for the current set of criteria. This count differs from the qualifying count which returns the total number of rows qualified when the rows found in the current search are merged with the rows in the qualified subset. This applies to oaqualify and the SQL extended command, QUALIFY. cursor A working area that defines the current position of a process in an open environment. In OmniAccess, cursors are established by a call to oaopencursor, and closed by a call to oaclosecursor. They are referenced through the cursor option as an integer value in calls to OmniAccess routines. In ODBC and JDBC, cursors are created with statement objects.

Contains Custom Indexes

data The material of a knowledge collection system used to create information The information stored in a database The representation of facts, texts, graphics, bitmap images, sound, video segments, etc…

data dictionary An information structure that defines the organization of a database including the list and description of each element of the database such as columns, data types, memory allocations for data storage, etc…

data files Files that contain actual data, rather than meta-data or other data descriptions. See meta-data and data.

data mart A departmental or subject-specific subset of data of a data warehouse. See data warehouse.

data mining The process of finding patterns or trends in a collection of data.

data modeling The development of rules that cover data entities (objects such as relationships, tables, columns, data integrity, etc.) of a database management system.

data set A collection of data entries where each entry contains a group of data items. Also called a table.

data source file An initialization file that stores the location of data on a server and the transport method used to access the data.

data source name (DSN) DSN; in ODBC, the name of the origin (e.g., of a file, database, document, etc.) of transmitted data.

data type The characteristic description of data, be it alphabetic, numeric, integer, decimal, character, string, etc.

data type discrepancy Condition in which data stored in a particular field does not match the DBMS data type defined for that field.

data warehouse A voluminous collection of summarized consistent data, usually historical in nature, used for only data analysis rather than data processing A database designed and optimized for OLAP.

database A collection of data files stored in a specific structure, whose organization is centrally defined in a root file or data dictionary.

database key A field or column that uniquely identifies a row in a table.

database tuning A process effected on a database by the database administrator (DBA) in an effort to improve the speed and efficiency of data retrievals and data throughput.

data-only option An option supported by the OmniAccess update routines (oainsert, oadelete, and oaupdate) that disables the real-time updating of OMNIDEX indexes. This option is useful when updating a high volume of records. Data-only updates must be followed by a DBINSTAL B or INDEX operation to update the OMNIDEX indexes.

DataSource a collection of information used by the ODBC software to connect to a database, required before accessing databases from a container application such as Microsoft Word.

DBA Database Administrator; person responsible for maintenance of a database system.

DBMS Database Management System; the file structure for the storage of data and the functions included in the maintenance of the data, such as transaction processing, table relationships, back-up and data restoration.

DCE Distributed Computing Environment; the term for the Open Software Foundation (OSF) standard of computer interoperability.

decompile in OMNIDEX, the process of reverting a compiled Environment Catalog file to an Environment source file using the utility OADECOMP. denormalization the process of introducing duplication of data within a relational database (which is typically normalized), used in data warehouse applications to avoid the inefficiencies of cross-table join operations. See normalization. detail data set Child table.

Detail Record a record in a child table. See DR Domain.

device driver a program that provides an interface between the operating system and a hardware device.

DI keys distributed index key. See distributed index.

dimension table A parent table with a unique key that provides keyed access into a child (fact) table; tables that contain descriptions for codes in fact tables.

direct access the process of locating a record or row based on its location on a storage device, which is fast but requires an exact key match. See hashed access and hashed key.

disconnect the process of terminating a network connection.

disk I/O See I/O.

distributed computing a hardware system architecture that spreads the computing load over multiple computers connected via a network.

distributed data the placement and maintenance of data at one or more data sites on a mainframe computer or across a telecommunications network.

distributed index An indexing option for Omnidex, or MDK indexes, in which the index is stored as a separate file, essentially creating an IMSAM, or ASK key; An ASK Index installed through an MDK Index process with ;DI option which can be used essentially as an MDK Index, including on-line index maintenance and optimization like an MDK Index, and also aggregation as on an ASK Index.

domain A parent table and any child tables that were linked (prejoined) to it during Omnidex installation; A single, unjoined table with Omnidex indexes installed on it; A a single data set installed with multiple keys; An association of several data sets installed with multiple keys and linked during installation. Also called an Omnidex domain. See SI domain and DR domain.

domain name server in TCP/IP, a server program that supplies name-to-address mapping by storing each network host's network address, domain name, and an optional alias.

DR See Detail Record.

DR domain an OMNIDEX domain that consists of one detail data set, or child table, that has not been linked to a master data set, or parent table. A DR domain child table can also be part of RS (Record Specific) domain. See Hybrid-DR domain and domain.

drill down In data analysis, the process of narrowing data to retrieve specific underlying detail; In OMNIDEX, successive iterations of adding criteria to a query, refining a search to the exact subset of required data.

Driver Manager Software written by Microsoft and installed when installing ODBC applications; Software accessed by container applications such as Excel which initiates the appropriate ODBC driver such as the OMNIDEX ODBC Driver.

DSN DataSource Name; see data source name

DSS Decision Support System; an application that provides information to a business analyst to identify organization options.

Domain

DR

EIS Executive Information System; an application that supports the strategic planning of an organization.

Environment a set of objects that can be manipulated to determine a user's access to data. An Environment determines what data is available to a user, what access a user has to that data, and so on.

Environment Catalog a file that contains information about how data is organized, what users can access it, and how they can access it.

exceptional conditions errors that occur after an intrinsic is called which prevent the call from executing successfully, for example, reading beyond the last key value in an ASK Index sequence.

excluded words list an ASCII file containing words to exclude from indexing. Also called an excluded words file.

excluded words words specified through OmniUtil for exclusion from the Omnidex indexes.

executable library XL; the location of the Omnidex intrinsics and call conversion procedures in native mode versions of Omnidex; the HPPA native mode library of callable routines.

extract specific data selected and retrieved from an existing database, usually from an OLTP system for use in data warehousing. See OLTP and data warehouse. SQL function that “extracts” specific parts of a date. ODXSQL function that creates SQL insert, update, and delete statements for a table to allow a user to restore a table to its original state.

extranet an Internet-based network where data is shared between one company and another, typically business partners.

Explain Plan

Full Text Index

fact table

• A single child table in a data warehouse containing potentially repetitive information, surrounded by parent tables (dimension tables) that are used to provide keyed retrievals of unique rows from the child table;
• The main table in a star schema containing metrics used for aggregations and codes that are linked to dimension tables, used to access descriptive information for the codes.

fat client

• See thick client.

File DSN

• in ODBC, a file level data source name.

flat file

• a file structure where data items are stored line by line in a two- dimensional array.

foreign key

• See child key

FTP

• File Transfer Protocol; a client/server protocol that allows a user on one computer to transfer files to and from another computer over a TCP/IP network.

• As a network term, a protocol server;
• As a Web term, an interface between an external source of information and a Web server.

• Omnidex intrinsics that are used to perform locks, updates, and other general functions.

• an Omnidex search by a partially specified key followed by an at sign ( @ ). Also known as a generic retrieval.

granularity

• a degree of detail data in a data warehouse where “high granularity” indicates a large quantity of detail with little summarization. “Low granularity” indicates a small quantity of detail, or data greatly reduced through summarization.

grouping

• an MDK key option (;Gn) that treats several multiple keys as one logical entity.

Hybrid DR

hashed access

    See direct access.

hashed key

• a database key facilitating direct access to data on a storage device. See key and direct access.

hashing

• an indexing technique where an algorithm is used to calculate the location of a row or record, requiring a full key look-up and an exact match, including upper and lower case characters, spaces, and punctuation.

having

• An SQL statement clause that allows additional criteria to be applied to an SQL statement after the WHERE and GROUP BY clauses have been processed. The HAVING clause is generally used to apply criteria to an aggregation as in “SUM(ORDERS.TOTAL)>10”.

high cardinality

• (1) High table cardinality indicates a large number of records in a table, a Sales Fact table for example.
• (2) High column cadinality indicates many possible unique values in a column. For example, names, addresses, account balances, product inventories, etc. See cardinality.

high-water mark

• See serial read.

hints

• a method of forcing the Omnidex Optimizer to use a particular retrieval method, in a star schema database environment, for example.

host

• a generic term for any device that allows communication from an outside source.

hosts file

• a file containing a cross reference between names for computers in a network and their physical network addresses, resident on the computer specified as the domain name server. See domain name server.

HPPA

• Hewlett Packard Precision Architecture; a RISC architecture used in series 9nn computers.

Hybrid DR

• Hybrid Detail Record; a variation of a multi-domain child, providing the ability to index a child table as a DR domain or in an SI domain. See DR domain and SI domain.

Internal ID list

I/O

• Input/Output; the physical process through which data is read from or written to a computer.

ID

• a 4-byte integer that is most often sequentially-ascending and which starts at 1, except when the Search Item (SI) is a 4-byte integer.

ill-behaved row ID

• unique record designations in a database table or set of data that are not neatly sequential; for example, the first record is identified as row ID 37, the second record in the series is row ID 43, the third record is numbered 11, and so on. See record number, row ID, and well-behaved row IDs.

index

• a list of key words or data items with associated pointers used to locate specific rows or records in a table of a database, facilitating fast access to data. See inverted, B-tree, and bitmap index.

indexing

• a method of creating file structures that have pointers into a database, facilitating directed access to rows or records in a table.

indexing discrepancy

• when the data in an indexed column is not indexed correctly in its OMNIDEX index. For example, if a native row is deleted, but the keywords indexed for it are not, and indexing discrepancy is said to exist.

indexing option

• an option installed on an OMNIDEX index that customizes how that index's key values (keys) are indexed. For example, the Soundex option causes keywords to be indexed for Soundex (phonetic) retrieval.

index-only mode

• term applied to high-performance retrievals and updates that affect the key value only. Discrete mode ASK Index retrievals return only the key values stored in the index sets without returning a record.

inner joins

• A join that retrieves rows from multiple tables by matching rows with common values in each table. Rows with no matching records in the joined tables are excluded.

instance

• identifies a unique connection by a user to an environment catalog.

internal ID list

• the most current list of record identifiers qualified by an ODXFIND.

intrinsics

• on an HP3000, the subroutines used to perform database functions, including the OMNIDEX, ODXFIND, and ODXGET retrieval functions.

inverted

• a style of indexing that uses a storage method opposite of that used in storing the data. To store data, a file starts with a record identifier, then associates data with it. Inverted indexing starts with a unique piece of data and associates record identifiers of records that contain the unique piece.

inverted list index

• a file structure that stores the pointers to a database as data, and the data from the database as keys.

ISAM

• Indexed Sequential Access Method; a type of database indexing through which data is retrieved by incrementally successive rows.

item list

• the items passed through the list parameter.

Java

• Programming language widely used for web programming developed by Sun Microsystems.

JDBC

• Java database connectivity standard defined by Sun Microsystems as an extension to the Java programming language.
• Omnidex provides its own JDBC driver to work specifically with the OmniAccess API.

join

• the process that interrelates the rows from two or more tables during an SQL Select look-up.

keep character

• when modifying a record in DATADEX, the character you type to leave the leading bytes of a character field unchanged so that you can modify the trailing bytes.

key

• a field, or column, that is used to select records from a data set, providing fast access to rows. Keys are typically indexed, uniquely identify parent rows, and provide a correlation for joining tables and maintaining data integrity across tables. See also Native keys and OMNIDEX indexes.

key options

• enhancements to OMNIDEX keys that are specified during installation to enable certain retrieval features (like Grouping) or update features (like Batch indexing).

keyed retrieval

• retrieval by an OMNIDEX or TurboIMAGE key field.

keyword

• A word or number that is indexed for an MDK Index
• An individual word or value in an MDK Index column delimited by spaces and punctuation. For example, if an MDK Index column contains “Kraft-Ebbing's Bait and Tackle” the words “KRAFT”, “EBBING”, “EBBING'S”, “BAIT”, and “TACKLE” are indexed as keywords.

keyword access

• the ability to find rows based on the words contained in a given field, regardless of the order in which the words appear. This is a primary feature of OMNIDEX indexes.

keyword indexes

• indexes composed of a word or words from a specified field in a database row. Renamed to MDK Indexes.

keyword lookup

• to look up data by words embedded within strings of text.

keyword retrieval

• refers to an OMNIDEX retrieval of record identifiers by any keyword value or combination of keyword values against an MDK Index.

keyword-only retrieval

• an OMNIDEX retrieval that fetches only keyword values from the indexes. Through OmniAccess, you would call oaqualify with the KEYWD option, and then call oafetchkeys.

left outer join

• An outer join in which all rows from the left table in the JOIN clause are included, only rows from the right table with matching rows in the left table are included, and left table rows with no matching right table rows are matched with nulls.

link field

• an Omnidex search by a partially specified key followed by an at sign ( @ ). Also known as a generic retrieval.

linked list

• a degree of detail data in a data warehouse where “high granularity” indicates a large quantity of detail with little summarization. “Low granularity” indicates a small quantity of detail, or data greatly reduced through summarization.

listening port

• an MDK key option (;Gn) that treats several multiple keys as one logical entity.

literal

• A letter, group of letters, or symbol that stands for itself as opposed to a feature, function, column value in a record.

logical link

• used in relation to Multifind, when data in two different data sets serve to create a correspondence between those records.

low cardinality

• Low table cardinality indicates few total records in a table.
• Low column cardinality indicates a low number of unique values in a column. For example, a “Yes/No” column has low cardinality because it has only two possible values.
• See cardinality.

MDK

MLH

Multi Level Hierarchies

MDD

• Multidimensional Database; a database architecture used to perform OLAP by summarizing data into a special database structure. See OLAP.

MDK Index

• Multi-dimensional Keyword Index.

meta-data

• also meta data; data which describes data, i.e., data characteristics such as field name, field length for data storage requirements, data type, field definition, data source, etc.

metrics

• Columns or fields that are used to summarize data;
• Columns that are used by the functions SUM, MIN, MAX, and AVG in container applications.

mode

• an intrinsic parameter that specifies what type of action is to be performed.

multi-dimensional

• a database design in which data is viewed using more than two sets of criteria.

Multi-dimensional Keyword Index

• composite (one or more items from a field) keyword keys composed of data from one or more tables. Also known as MDK Indexes. See OMNIDEX indexes and keyword indexes.

Multifind

• a retrieval feature that supports keyword searches across domains using field values from previously qualified records as keyword arguments against a table in a different domain.

Multiple key

• a field specified for OMNIDEX keyword retrieval. See OMNIDEX indexes, keyword indexes and MDK Index.

Multiple key access

• the ability of OMNIDEX to qualify records based on arguments entered against several keys.

Multi-RIN (MR) capability

• a feature used to lock several data resources simultaneously. This capability is used by programs that update or lock OMNIDEX databases, data sets, or items.

native key

• a key (index) defined for a database or system of files before OMNIDEX was installed on it. OMNIDEX uses native keys when prejoining tables.

network transport

• the method that OMNIDEX Client uses to transfer messages between the client and the server process.

next free ID

• an internally maintained value that indicates the next assignable integer SI value. May be an ID previously used by a deleted record.

node

• a computer, device, or other entity that is a part of a network.

noise words

• keywords that are useless for retrieval, either because they have little meaning, or are used too often to distinguish rows from one another. For example, “the”, “a”, and “an” are noise words.

normalization

• the process of reducing or eliminating duplicate data within a database for the purpose of restructuring data relationships to the most simple form.

object

• an entity within the Environment that can be accessed and manipulated, a column definition, for example.

ODBC

• Open Database Connectivity; (1) a standard for accessing different databases from PCs or other small workstations, providing a common SQL-based method for translating requests into the database tool's native SQL (2) software developed by Microsoft that enables Windows applications to connect to various databases.

ODBC Administrator

• software written by both Microsoft and the ODBC Driver software company to create and maintain DataSources.

ODBC Driver

• software written by DISC specifically for Omnidex-enhanced databases;
• the Omnidex module that sends ODBC calls to the Omnidex retrieval engine that accesses the database, attached to a container application by Driver Manager.

ODBC-aware applications

• an application that uses Microsoft's ODBC technology to access databases.

ODS

• Operational Data Store; an integrated collection of data that is subject- oriented and volatile, designed specifically for daily operations analysis.

OLAP

• On-line Analytical Processing; a method of data analysis used in data warehouses involving viewing data through multiple dimensions, aggregations, relationships, patterns, trends, and exceptions.

OLE

• Object Linking and Embedding; Microsoft Corporation's API for data access that manipulates data as objects and which supports Microsoft Corporation's Distributed Component Object Model (COM).

OLTP

• On-line Transactional Processing; routine recording and analysis of daily business activity, e.g., order processing, accounting, manufacturing, etc.; also known as operational processing.

Omnidex

• an advanced indexing product from Dynamic Information Systems Corporation (DISC), providing greater query flexibility than native database methods and accelerating database search and retrieval by reducing I/O.

Omnidex Access

• the Omnidex open systems API that performs the index look-ups, qualifying counts, and data retrievals, including the routines oaqualify, oaselect, and oafetch. Also known as OmniAccess.

Omnidex condition word

• word 11 of the status array, which contains an error code if a call to an Omnidex intrinsic failed.

Omnidex detail

• a detail set linked to an Omnidex master by an Omnidex SI (search item).

Omnidex domain

• see Domain.

Omnidex error indicator

• the word of the status array that indicates whether an error has occurred. This word differs between default (active) error handling and passive error handling.

Omnidex ID

• a binary integer (I2, J2, or K2) value that is used to identify records during keyword searches (for Multiple keys);
• the arbitrary sequential ID of a record assigned by Omnidex.
• Also known as ODXID or ODX ID

Omnidex indexes

• the indexed columns or composite indexes designated for Omnidex keyword or sorted indexing through the DBINSTAL installation utility program.

Omnidex Intrinsic Interface

• Omnidex intrinsics that provide retrieval access to Multiple and Sorted key indexes and maintenance to Omnidex indexes.

Omnidex master

• a master data set that contains one or more fields specified for Omnidex keyword retrieval, or a master that is linked by the Omnidex SI to a detail that contains one or more Multiple keys.

Omnidex root file

• the index file that contains information about the database structure, including which fields are specified for Omnidex retrieval.

Omnidex Search Item

• the search item field in an Omnidex master which may also be the Omnidex ID for the set (that is, the SI-ID). See Search Item.

one-to-many relationship

• refers to data entities where one entity corresponds to many other entities, as in the case of a search item value to many detail records.

one-to-one relationship

• refers to data entities where one entity corresponds to only one other entity, as in the case of a search item value to a master record.

operator

• any special character or token used in Omnidex retrievals, for example, the Multifind operator (&).

OR criteria

• Multiple criteria values or predicates that records must match one or more of, but not all.

outer join

• A join that includes all rows from joined tables that meet the specified search criteria. Rows without matching records in the joined table are matched with null values.

OSF

• Open Software Foundation;
• a foundation created by nine computer vendors;
• the computer operating system developed by the OSF.

Partitioned Databases

Partitioned Tables

Powersearch

Prejoined Indexes

Prejoned Tables

parent key

• the column within a parent table that uniquely identifies rows or records containing the same values.
• Also known as Primary Key.
• See also: key

parent table

• a table where any row is uniquely identifiable by the contents of a single field (a key); Parent tables are often superior to child tables, typically in a one-to-many relationship.

parsing

• separating a character string into individual keywords.

partial-key retrieval

• refers to an ASK Index or MDK Index retrieval where a partially specified search argument is used to retrieve one or more records.

partitioning

• the dividing or grouping of data in a database across storage media such as a disk drive, or across nodes in a distributed computer or network environment in order to speed processes by concurrently processing I/O.

path

• information included with a file name that tells the location of the file. How you specify a path is dependent on the operating system. In MPE, a path is filename.group.account. In VAX/VMS, a path is node::device:[directory.subdirectory.subdirectory…]filename. In UNIX, a path is directory/subdirectory/subdirectory/…filename.

pattern matching

• specifying wildcard notation, for example, asterisks (*), pound signs (#), percent symbols(%), and so on, in a value which represents single or multiple characters in that value.

port

• in a computer network, an identifying number that supports transport between a server and one of several clients.

prejoined tables

• tables that were linked together during OMNIDEX installation using KEY values declared in the OMNIDEX environment catalog. Prejoining tables optimizes OMNIDEX join operations between them.

progressively qualifying

• performing several keyword searches in succession against the same table to narrow the qualified subset of rows. Also called a “drill-down” search.

qualified subset

• the list of rows established and refined by searches of keyword keys. You can either retrieve the rows represented in the qualified subset by calling oaselect and oafetch, or you can further qualify the rows by successive calls to oaqualify.

qualify

• Omnidex Extended SQL statement that “qualifies” rows based on specifed criteria.
• Obtain an Omnidex ID list of records that match specified criteria.

qualifying count

• The number of rows currently represented by the qualified subset. This number is returned to the oaqualify status structure and is displayed at the bottom of the prompt screen for an OMNIDEX operation in DATADEX.

query

• a request made to a database for a subset of records defined by specific criteria.

RC Record Complex - a type of Omnidex Domain where child records are collectively indexed together with the parent record.

RC child a child table which is indexed with only primary keys (from parent table). See Record Complex.

RDBMS Relational Database Management System; a database approach involving two-dimensional tables to define data relationships.

record A set of field, or column, values in a data file or database table; (2) a structured collection of information or data treated as a unit; also known as a row or tuple.

Record Complex when a child table is prejoined (linked) to a parent table, an affiliation of records in an SI domain that all contain the same OMNIDEX SI value; for example, qualifying from child or parent returns parent rows; keys that qualify record complexes (MDK Indexes in parent tables or master data sets, and MDK Indexes installed with the Record Complex option). When Record Complex is installed on a keyword index in a child table, the index is treated as if it belonged to the parent table. See SI domain, Record Specific, RC child, RS child.

record identifier A value that uniquely identifies a record in a table; A search item (SI) for a parent table and a row ID for all other tables.

record number an internally maintained value associated with the position in a detail file of individual detail records.

Record Specific - A type of Omnidex Domain that indexes child records individually instead of collectively.

Record Specific Table, an affiliation of records in an SI domain that are indexed in reference to their own table; for example, if qualifying records from a child table, query will return child rows with the ability to access the parent rows, else if qualifying from the parent table, query will return parent rows; refers to MDK Index or keyword retrievals that qualify individual detail records. Non-Record Complex keys in detail records are Record Specific. See Record Complex and SI domain.

record specific ID cross reference a portion of an OMNIDEX index that correlates each parent constraint with the corresponding child row IDs. This portion of the index is used for performing qualifications, as well as several types of joins.

redefineable row IDs row IDs subject to OMNIDEX manipulation to insure sequential nature. See well-behaved row IDs.

Referential Integrity A method of ensuring that the relationship between records in related tables is maintained. Specifically, referential integrity checks to make sure that a parent record exists for every child record.

relational operator the greater than (>), greater than or equal to (>, less than (<), less than or equal to (⇐), and equal ( operators used to specify a relational retrieval. relative record number see record number.

repeating key See child key.

replacement step the process of replacing ID values in the kernel aggregate results with the column specified in the group by the statement. The ID value is used to read into the dimension or snowflake table where the Group By column resides.

retrieval intrinsics intrinsics that are used to search for or retrieve records.

right outer join An outer join in which all rows from the right table in the JOIN clause are included, only rows from the left table with matching rows in the right table are included, and right table rows with no matching left table rows are matched with nulls. When Omnidex encounters a right outer join, it will attempt to convert it to a left outer join by switching the tables. If unable to do this, the query will fail.

RMS Record Management Services; the VAX-based database management system software from Digital Equipment Company (DEC).

ROLAP Relational On-line Analytical Processing; on-line analytical processing using summary tables in a relational database. See OLAP. MDD, and MOLAP.

Rollup Table

root file a file that contains information about the database structure. In OMNIDEX, the OMNIDEX root file contains information about where keys are installed.

row See record.

row ID a unique representation of a single record in a set of data. See record number.

RS child a child table indexed with both the primary keys (from parent) and its own row IDs. See Record Specific.

run-time options configurable options that control certain features of a program, like DBINSTAL, for the duration of a session.

samelist

• Refers to a previously qualified list of Omnidex ids. Used in multiple qualifies to refine queries.

samelist operators

• AND, NOT, AND NOT (AND_NOT in oaqualify)
• Included in a QUALIFY statements WHERE clause preceding the criteria predicate
• Specified as an option is OAQUALIFY

scalability

• The ability of a system to easily adapt to support larger or smaller volumes of data and more or fewer users.

schema

• Database layout. A detailed and possibly graphical representation of a database's design.

scrubbing

• The process of cleaning data to make it consistent in preparation for data warehouse retrieval.

search item

• Primary key - SI
• A field within a table that is used for calculated access in parent tables, and chained access in child tables

select item

• An item in a SQL SELECT statement's select-list. See the SQL Reference for details on valid select-list items.

select-list

• A comma separated list of items to be returned to the application in a SQL SELECT statement. See the SQL Reference for details on valid select-list items.

sequential read

• A search of a file or database performed by comparing successive records with search criteria.

serial read

• A sequential read of records in a data set that is performed one block of records at a time.

server

• A hardware component dedicated to a specific purpose.
• The part of the client/server architecture which refers to the host CPU where the database and other corporate information is stored.

SI domain

• An affiliation of one parent table, and one or more child tables, linked at installation, whose keyword values are all indexed in the same indexes. See also: search item.

snowflake schema

• A type of database design used in data warehousing involving multiple tables in a hierarchy within a given dimension i.e., data is organized into parent tables with branches to child tables and so-called grandparent tables (snow flakes).
• When the schema is displayed graphically, it has the appearance of a snow flake.

soundex

• ;SX
• an indexing option which allows data retrieval based on phonetic arguments, useful for location of data with keys that are difficult to spell or that can be spelled in several ways.

source

• In Multifind usage, refers to the item, set, or database that is supplying the data values used as search arguments.
• Textual, uncompiled version of an Omnidex Environment Catalog.
• Textual, uncompiled version of a computer program.

special characters

• A character that is parsed by Omnidex before being indexed
• A character that is interpreted by Omnidex to have special meaning

SQL

• Structured Query Language
• A standardized computer language used to add, modify, delete, and retrieve information from a database

star schema

• A type of database design used in data warehousing composed of a large sized fact table combined with numerous dimension tables, i.e., parent tables branched to child tables.

summary table

• A table containing pre-summarized data, typically used in ROLAP data warehousing, used to perform fast look-ups against specific, predictable combinations of criteria.
• Omnidex aggregation indexes eliminate the need for summary tables.

superview

• The technique of flatening out multiple tables into a single table to facilitate very high-speed Omnidex access.

synchronous

• A process that is executed in conjunction with another process
• In Omnidex, index updates that are update at the same time that the data updates occur.

System Data Source

• Windows only. An ODBC data source that is stored in the System's registry. An Omnidex data source can be a System Data Source.

table

• a collection of data typically organized in two or more dimensions, usually by row and column (or field).

table cardinality

• The number of rows in a table. See also: column cardinality, cardinality.

table scan

• an exhaustive read of all rows of data in a table, comparing search criteria against the respective columns, and which can involve significant amounts of I/O activity.

target

• In Multifind, refers to the domain, data set, or field in which a list of keyword values is to be further qualified.
• Regarding DISC utilities, refers to a data entity on which a specified operation is to be performed.

TCP Sockets

• one of the transport methods used by OMNIDEX Client for transferring data packets between the client and server, used when clients access the server through a direct connection, as through a LAN.

TCP/IP

• Transmission Control Protocol over Internet Protocol; the de facto standard Ethernet protocol used for inter-networks in the UNIX environment.

textual data

• data containing words, phrases, and sentences (as opposed to numeric or other fixed data).

thick client

• a client program or device which contains most of the presentation layer and program logic. See also: tier, thin client.

thin client

• a simple client program or device which relies on the server for most of the function of the system, are easy to maintain and deploy since program modifications are made only in a centralized location, and installation of updates are not required on each client seat. See also: tier, thick client.

tidmap

• tuple identification map; in OMNIDEX, six digits separated by commas that define the translation of native database row IDs to OMNIDEX IDs.
• As of version 4.0, tidmaps are no longer used.

tier

• a level describing client/server architecture, where a two-tier model is a thick client model (e.g., a PC running a user interface as client, combined with a host/server on which a database and the application resides). A three-tier model is a thin client model (e.g., a PC with a user interface as a client, a separate server contains a database, and a third server contains an application program).

TR - Transparency

• in OMNIDEX, the correlation of native database row IDs with OMNIDEX's internal equivalent through the use of a cross reference file.

TRR - Transparency / Row Number

• Transparency with Record Number; a variant of TR; used on index installation on flat file databases, relational databases with composite search items, and when it is necessary to intersect aggregation indexes with MDK qualifications in a TR parent table.

transformation

• as a data warehouse term, the data conversion process involved in transferring data from a legacy system to a data warehouse system.

transport

• a setting in the data source file that shows which transport method to use. See TCP Sockets and WRQ PPL.

trigger

• a process or set of instructions that is activated automatically within a DBMS based on certain events. For instance, an update trigger would automatically run everytime a database table was updated.
• used by ODXAIM to update the Omnidex indexes.

TSE

• Technical Services Engineer; a representative of DISC who provides OMNIDEX troubleshooting expertise, and who provides on-line and telephone support to customers. TSEs do not address database design issues; PSEs provide database design and OMNIDEX implementation consulting services.

tuning

• A process effected on a database by the database administrator (DBA) in an effort to improve the speed and efficiency of data retrievals and data throughput. Also known as database tuning.

tuple

• A set of field, or column, values in a data file or database table; (2) a structured collection of information or data treated as a unit; also known as a row or record.

unique key

• any column (preferably indexed) that uniquely identifies individual rows in a parent table.

unlinked child tables

• child tables that are not prejoined to a parent table during OMNIDEX installation.

upper case

• Alphabetic characters in capital form. ie, A is upper case, a is lower case. See also: lower case.

upshift

• Shift all alphabetic character to upper case.

user DSN

• An ODBC user level machine registry data source object.

wall time

• elapsed time as clock time or real time, as contrasted with CPU or system resource time.

well-behaved rowids

• unique record designations in a database table or set of data that are neatly sequential; for example, the first record is identified as row ID 1, the second record in the series is row ID 2, the third record is numbered 3, and so on. See record number, row ID, and ill-behaved row IDs.

wild card

• any character that is used to imply a partial search argument. The at sign ( @ ) is used as a wildcard in generic retrievals.