Glossary

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.

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.

^ Index ^| A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z | $ |