Glossary: Glossary :: SAS/Warehouse Administrator(R) 2.3 Metadata API Reference, Second Edition

Glossary

application program interface (API) interpreter

specifies a program that translates the metadata type that is requested by a client to the corresponding metadata object in a repository.

business metadata

represents text that describes the content or purpose of an application element. For example, the business metadata for a SAS table might describe the purpose of the table and contact information for the person who is responsible for the accuracy of the information in the table.

component

indicates a set of related metadata types. Each component has an ID, such as WHOUSE, and a name, such as SAS/Warehouse Administrator, that often matches the name of the application whose metadata is modeled by the component. WHOUSE is the component that is used for SAS/Warehouse Administrator.

contact records

indicates a metadata record that specifies the owner or administrator who is responsible for a given warehouse element. An owner is a person who formulates policy and makes decisions about an object. An administrator is a person who implements decisions that are formulated by the owner in accordance with established policy.

You can include contact records in the metadata for groups, data stores, processes, jobs, and other objects in the current environment. They are part of the global metadata for a warehouse environment.

Although contact records are not required for code generation, you might find them essential for project management. They enable you to identify--and perhaps programmatically contact--the people who are responsible for a given warehouse element.

data file

represents a metadata record that specifies a SAS file that is an input to an ODD.

If you are defining an ODD whose Load Step is a DATA step view or an SQL view (but not a Pass-Through view), you must define its inputs in the Process Editor. Even if your ODD does not meet the conditions above, you might still want to specify a process flow for this job for documentation purposes.

You can define a data file that simply registers the location of a SAS table or view or are that registers the location of a DBMS table with the help of a DBMS LIBNAME definition. You can also define a data file that extracts information from a data source, saves the results to a SAS table or view, and then specifies the location of the extraction table or view.

Data files are added in the Process View of the Process Editor. In the process flow for an ODD, you can add a data file by clicking the ODD (or the background) with the right mouse button, selecting Add, and then New data file.

data group

specifies a simple grouping element for data tables, InfoMarts, and other data groups.

In the SAS/Warehouse Administrator Explorer, you can add a data group only to a subject, a data warehouse, or another data group.

You could use a data group to implement a data mart.

data table

represents a metadata record that specifies a SAS table or view or a DBMS table or view that can serve multiple purposes. Data tables are frequently used to define intermediate data stores, such as look-up tables that are included as part of a join. You can use them to define detail data stores, summary data stores (if you write your own summary code and register it as the Load Step for the data table), or tables that hold information that does not fit anywhere else.

In the SAS/Warehouse Administrator Explorer, you can add a data table to a data group.

data transfer process

specifies a metadata record that is used to generate or retrieve a routine that moves data from one host to another. Data transfers are required when an input source and the target data reside on different hosts.

If SAS/Warehouse Administrator generates the code for a data transfer, it uses SAS/CONNECT software and PROC UPLOAD or PROC DOWNLOAD to move the data. This method is most easily applied to transfers between a local host (host where SAS/Warehouse Administrator is installed) and a remote host.

If you need a remote-to-remote transfer, one solution is to specify a user-written transfer routine in the metadata for the data transfer process. The SAS/Warehouse Administrator User's Guide might offer other solutions for remote-to-remote data transfers.

Note: Data transfers always execute on the remote host (a host other than the host where SAS/Warehouse Administrator is installed). [cautionend]

A data transfer, like a mapping, a user exit, or record selector, is inserted in the process flow for a data store.

data warehouse

represents a metadata record that specifies the SAS library _DWMD. The _DWMD library is the metadata repository for most groups and data stores in a data warehouse or a data mart at your site.

In the SAS/Warehouse Administrator Explorer, you can add a data warehouse object only to a data warehouse environment.

DBMS connection profile

represents a metadata record that specifies a user name, a password, DBMS options, and other information that SAS can use to access source data or warehouse data stores in a database management system (DBMS) other than SAS. DBMS connection profiles are included in the metadata records for DBMS data stores or DBMS LIBNAME definitions in the current environment. DBMS connection profiles are part of the global metadata for a warehouse environment.

If you want SAS/Warehouse Administrator to generate code that will access source data in a DBMS or load warehouse data in a DBMS, you will probably need at least one DBMS connection profile for each target DBMS.

If you want to connect to the same DBMS but with different levels of privilege or with different options, you need to create different DBMS connection profiles with the appropriate user names, passwords, and options.

DBMS LIBNAME definition

specifies a special SAS library definition that you can use to extract source data in DBMS format or to create warehouse data stores in a DBMS.

SAS/Warehouse Administrator uses a DBMS LIBNAME definition to generate a SAS/ACCESS LIBNAME statement. Some of the metadata that you specify in the definition corresponds to the options in the LIBNAME statement. For example, a DBMS LIBNAME definition specifies a SAS/ACCESS engine--such as Oracle or SYBASE--that enables you to access the corresponding DBMS as if it were a SAS library.

A DBMS LIBNAME definition also specifies a DBMS connection profile, which includes the DBMS user ID, password, server name, and other connection information that is used to access the DBMS. These options are passed to DBMS client software, which makes the connection to the DBMS.

DBMS LIBNAME definitions are part of the global metadata for a warehouse environment. You can use DBMS LIBNAME definitions in ODDs to access source data in DBMS format. By default, for new DBMS data stores, SAS/Warehouse Administrator generates Load Steps that use SAS/ACCESS LIBNAME statements.

For details about the SAS/ACCESS LIBNAME statement, see SAS Language Reference: Dictionary.

detail logical table

represents a metadata record that specifies a SAS table or view that can serve multiple purposes. A detail logical table is often used to implement a view on multiple, related detail tables. You can use detail logical tables as grouping elements for detail tables or as detail data stores.

In the SAS/Warehouse Administrator Explorer, you can add a detail logical table only to a subject. A subject can have only one detail logical table. A detail logical table can contain any number of detail tables. Detail logical tables in different subjects can share (link to) the same detail table.

detail table

indicates a metadata record that specifies a SAS table or view or a DBMS table or view that serves as a detail data store.

In the SAS/Warehouse Administrator Explorer, you can add a detail table only to a detail logical table. A detail logical table is often used to implement a view on multiple related detail tables.

event

is a metadata record that specifies a condition for controlling a job, such as checking for certain return codes or verifying the existence of a file. To use events, you must create them, include them in a job flow, then write a metadata API program that reads the job flow and generates code for it.

You can add a new event only in the Job Hierarchy view in the Process Editor. In the Job Hierarchy view, an event can only be added to a data warehouse environment, data warehouse, or a job group.

external file

specifies an input to an ODD that extracts information from one or more sources that are not in SAS format. That is, an external file is an input to an ODD whose Load Step is a DATA step view.

External files are added in the Process View of the Process Editor. In the process flow for an ODD, you can add an external file by clicking the ODD (or the background) with the right mouse button, selecting Add, and then New External File.

host definition

indicates a metadata record that specifies a computer where data stores reside, where processes and jobs execute, or where process output is sent. Host definitions are included in the metadata records for data stores, processes, and scheduling server definitions in the current environment. Host definitions are part of the global metadata for a warehouse environment.

Host definitions are required in order to access source data and to load warehouse data stores.

InfoMart

(also called an information mart) specifies a simple grouping element for InfoMart items and InfoMart files.

Unlike most objects in SAS/Warehouse Administrator, InfoMart items and InfoMart files are used to display information rather than store it. For example, you might use an InfoMart item to display a chart that summarizes sales information from a warehouse data store. Also you might use an InfoMart file to open a spreadsheet that contains information that is useful to the person who is managing a given warehouse environment.

In the SAS/Warehouse Administrator Explorer, you can add an InfoMart only to a subject, a data group, or an ODD group.

InfoMart file

(also called an information mart file) indicates a metadata record that specifies a file other than a SAS file that you want to register in a warehouse environment. The file can be a spreadsheet, an HTML report, or any file that you can open using an external application. InfoMart file metadata describes the location of an external file and the technique for opening that file.

Unlike most objects in SAS/Warehouse Administrator, InfoMart files are used to display information rather than store it. For example, you might use an InfoMart file to open a spreadsheet that contains information that is useful to the person who is managing a given warehouse environment.

In the SAS/Warehouse Administrator Explorer, you can add an InfoMart file to an InfoMart.

InfoMart item

(also called an information mart item) indicates a metadata record that specifies a routine that generates output from detail data stores or summary data stores in a data warehouse. The output is usually a SAS chart, report, graph, or query result.

Unlike most objects in SAS/Warehouse Administrator, InfoMart items are used to display information rather than store it. For example, you might use an InfoMart item to display a chart that summarizes sales information from a warehouse data store.

In the SAS/Warehouse Administrator Explorer, you can add an InfoMart item only to an InfoMart.

job

indicates a metadata record that specifies the processes that create one or more data stores (output tables).

A job must include a process flow if SAS/Warehouse Administrator will generate the source code for the job. If you will supply the source code for a job, no process flow is required, but you might want to create one for documentation purposes.

A job might include scheduling metadata that enables the process flow or user-supplied program to be executed in batch mode at a specified date and time. A job might also include a job flow.

load process

specifies a metadata record that is used to generate or retrieve a routine that puts data into a specified target object. After you define the metadata for a given data store, you must define a load process, which creates and loads the data store.

To define a load process for a given data store, display that data store in the Process View of the Process Editor, click its icon with the right mouse button, and select Edit Load Step.

metadata

specifies information that is internal to an application that describes elements in the application, such as tables and columns. There are two main kinds of metadata: physical metadata and business metadata. See also business metadata and physical metadata.

Most SAS/Warehouse Administrator metadata contains information about data sources, data stores, and the jobs that extract, transform, and load source data into the warehouse data stores. SAS/Warehouse Administrator metadata is stored in two or more metadata repositories.

metadata application program interface (API)

specifies a set of software tools that enable programmers to write applications that access metadata. The SAS/Warehouse Administrator metadata API enables you to access metadata in SAS/Warehouse Administrator.

metadata client

indicates an application that uses metadata API methods to read or write metadata. For the current release of the SAS metadata API, metadata clients must be written in SCL.

metadata object

represents an instance of a metadata type--the metadata for an element in an application, such as a table or column.

metadata property list

specifies a list of properties for a given metadata type that you pass to a metadata API method or a list of properties that a metadata API method returns to you.

metadata repository

indicates a data store that contains an application's metadata.

metadata type

represents a template that models the metadata for a particular kind of object in an application. The parameter list for a metadata type matches the items of metadata that are maintained for the corresponding object.

ODD

(operational data definition) specifies a metadata record that provides access to data sources. The ODDs, in turn, are used as inputs to detail data stores in a warehouse environment.

At a minimum, in order for a data source to be visible in a warehouse environment, you must specify the location of that data source in an ODD. You can define an ODD that simply registers the location of a SAS table or view or one that registers the location of a DBMS table with the help of a DBMS LIBNAME definition. You can also define an ODD that extracts information from a data source, saves the results to a SAS table or view, and then specifies the location of the extraction table or view.

In the SAS/Warehouse Administrator Explorer, you can add an ODD only to an ODD Group.

ODD group

specifies a simple grouping element for ODDs. It might also contain InfoMarts.

In the SAS/Warehouse Administrator Explorer, you can add an ODD group only to a warehouse environment.

OLAP group

(online analytical processing group) organizes related summary data, which is stored in OLAP tables or OLAP MDDBs. The OLAP group properties specify the logical structure of the summarized data and how they relate to the detail data in a data warehouse. OLAP groups have a type attribute, which you specify as: ROLAP, MOLAP, HOLAP, or MIXED.

In the SAS/Warehouse Administrator Explorer, you can add an OLAP group only to a subject.

OLAP MDDB

indicates a metadata record that specifies a SAS MDDB. A SAS MDDB is not a SAS table. It is a specialized storage format that stores derived summary data in a multidimensional form, which is a highly indexed and compressed format. To load an OLAP MDDB, SAS/Warehouse Administrator generates code for the MDDB procedure, which summarizes data similar to the SUMMARY procedure.

OLAP MDDBs are the only kind of data stores in an OLAP group of type MOLAP. You can include OLAP MDDBs in an OLAP group of type HOLAP or MIXED.

Each MDDB in an OLAP group of type MOLAP must have an NWAY crossing that represents all of the data summarized to the lowest level, and it must be named NWAY. The MDDB can also contain additional crossings.

In the SAS/Warehouse Administrator Explorer, you can add an OLAP MDDB only to an OLAP group.

OLAP table

indicates a metadata record that specifies a file to store derived summary data. This file can be a SAS table or view or a DBMS table or view. An OLAP table can have multiple crossings.

To load an OLAP table, SAS/Warehouse Administrator generates code for the SUMMARY procedure, which summarizes data by computing descriptive statistics for columns across rows or within groups of rows.

OLAP tables are the only kind of tables in an OLAP group of type ROLAP. You can include OLAP tables in an OLAP group of type HOLAP or MIXED.

In the SAS/Warehouse Administrator Explorer, you can add an OLAP table only to an OLAP group.

physical metadata

specifies a set of software instructions that describes an application element. For example, the physical metadata for a SAS table might specify a certain number of rows and columns, with certain transformations applied to some of the columns.

process

specifies a metadata record that is used to generate or retrieve a routine that creates warehouse data stores extracts data, transforms data, or loads data into data stores. Mappings, user exits, data transfers, record selectors, and load steps are all processes.

Each process that you define in the Process View of the Process Editor generates or retrieves code. SAS/Warehouse Administrator can generate source code for any process except a user exit or an ODD load step. However, you can specify a user-written routine for any process.

record selector process

specifies a metadata record that is used to generate or retrieve a routine that subsets data prior to loading it to a specified table.

Note: In the current release, you can use a record selector only to subset the source data that is specified in an ODD or in a data file (which is an input to an ODD). [cautionend]

A record selector process, like a mapping process, a user exit process, or a data transfer process, is inserted in the process flow for a data store.

SAS library definition

specifies a metadata record for a SAS library that contains data, views, source code, or other information that is used in the current warehouse environment. SAS library definitions are included in the metadata records for data stores, processes, and jobs in the current environment. Library definitions are part of the global metadata for a warehouse environment.

Library definitions are required in order to access source data and to load warehouse data stores.

Note: A SAS library definition does not include a host definition. In a separate task, you must create a host definition for the host where the library will reside. In the metadata for data stores and other objects, you must specify both the library definition and the host definition for the computer where the library resides. [cautionend]