 |
 |
•
Before I get started, what should I know about Database objects?
• Before I get started, what should I know about
Database Access?
• Where do I get the ODBC database drivers from?
• Before I get started, who should I talk to about
Database permissions?
• Will I see only those business elements that I have
permissions for?
• What does “metadata” mean?
• What is the difference between a “Search”
and a “Query”?
• What types of databases can and cannot I search and
query?
• Can I search multiple databases at once?
• How do I get to a query after a search?
• Do I always need to search before I can query?
• How fast can I expect a search to be?
• How can I speed up my search?
• After a search, how do I know if a result is what
I’m looking for?
• How can I see information related to my data?
• Can I save the results of a search and a query?
• What types of SQL queries can I perform?
• Can I save SQL and open SQL text?
• How can I see the results of my query?
• How can I view multiple tables and queries at once?
• How can I create a local copy of a database?
Databases are the generic term for a structure of data controlled
by a database management system and organized generally (except for desktop
databases like ACCESS ) according to the following general-to-specific
structure:
Server -> Database -> Schema/Owner -> Table/View -> Column
Servers are usually equated with a physical computer (mainframe, mini,
or workstation) but actually more than one database server may reside
on the same computer. If your company has an e-commerce website
using Oracle, the database server may be called “Oracle_Web”.
The server may contain two databases: a production database and a testing
database (where changes are tested out before rolling out). So inside
of the Oracle_Webserver is a “Oracle_Prod” database and an
“Oracle_Test” database. Now if your company, say, has
two major subsidiaries (e.g. Commodities and Credit), each of these may
have their own schema inside of the Oracle databases, “Commodities”
and “Credit” (schemas are also sometimes called “owners”
or “creators”). Each of these schemas may have their
own “Customer” and “Account” tables or other business
tables that contain columns such as “Name”, “Address”
or “Balance”. MetaTrieve will search all servers, databases,
schemas, tables and columns as determined by the options on the administration
and search pages. MetaTrieve will also search views (which are defined
subsets or joins of tables) which can have their own unique column names
that may match the business semantic searched for even if the source table/column
does not.
MetaTrieve is an enterprise tool; this means that to use the
search and query features on any but desktop databases requires the drivers,
setup and configuration of the network databases in the company.
You will need the following to use MetaTrieve on remote databases:
1. Network and domain access to the database server desired.
The network administrator can provide this if the server does not show
in “Network Neighborhood” of Windows Explorer.
2. Database ODBC drivers for the database vendors (providers) used
in the tool.
3. ODBC (Open Database Connectivity) setup. You will need
to create a DSN (Data Source Name) for each database server you wish to
use in MetaTrieve (if not already on your system). For Windows,
this is usually a System DSN and for Windows XP, a User DSN. (For
ACCESS, this is always a User DSN and is probably already present--you
will only need to assign a default database on the configuration page.)
To configure a DSN, you will need the server name, Userid and password
(and for some databases, a database name--some default). Your administrator
can supply you with these items.
While the drivers for Microsoft SQL Server and other databases
may already be on your workstation from your install of Windows, Excel,
Access or Word, other vendor drivers (IBM’s DB2, Oracle, Sybase,
etc.) can be downloaded from the provider websites such as www.microsoft.com,www.oracle.com,
www.sybase.com, www.MySQL.com and www.ibm.com. Third-party vendors
such as StarQuest also provide database drivers. You can view your
database drivers on your computer via the “Drivers” tab on
the ODBC panel off of the Windows Control Panel or the “Administrative
Tools” in Windows XP.
You should talk to your database administrator about database
permissions. You will need to acquire a Userid and a Password from
the database administrator for each database you need to search and/or
query. Database permissions generally involve read (query), insert,
update and delete permissions for database schemas and tables. While
MetaTrieve itself will allow you to execute queries that involve all of
these, the database security system may prevent you from any of these
activities.
You will be able to search database metadata and view the results
(database names, table names, column names and column comments/remarks)
but you will not be able to see entire table descriptions, sample data
or queried data without read permissions.
Metadata is a generic term for “data about data”
or “descriptions of data.” It is primarily the description
of the business category of the data: “Joe Williams”
is a customer name, “Chicago” is a city name, Z12E34536 is
a purchase order number, etc. This category definition in databases
in normally contained in the labels of tables and columns, sometimes divided
between the two (the table may be labeled “Customer” and the
column “Name” or “FirstNm” and “LstNm”,
for example).
Finding these categories of data is what is normally referred to as “locating
the data” in order to query or process. “Discovery”
of data usually means determining what categories of data are in a database
or databases and in what database, table, and column the category is defined.
In web documents and XML, the metadata descriptions or categories (if
present at all) are wrapped around the data in the same document and you
can search on a city name (e.g. Atlanta) or on the term “city name”
and hope to get relevant information back; in databases, you generally
cannot get to the actual data (“Atlanta”) without first locating
the metadata or placeholder category (“city name”) that pertains
to the information desired. Then you can query or report on the
data (“Where are all my company stores in Atlanta?”).
A “Search” is a scan of database metadata for tables
and columns that may match your business search term. It finds the
“location” (database name, table name and column name) of
the data. A “Query” is an SQL request for the actual
data from a database, table(s) and column(s) that was located with the
metadata search. An example of a MetaTrieve search is simply a business
term “product price” entered that returns, say, Database:
eCommerce; Table: “Catalog”; Column: “Prod_Price”.
An example of a query after the search has found the database objects
is: “Select Product_Cd, Prod_Price from Catalog where Prod_Price
> 129.00”. First you search, then you query.
You can search and query any ODBC compatible relational database
(e.g. SQL Server, DB2/UDB, Oracle, MySQL, Sybase, etc.). You cannot
search and query legacy databases such as IBM’s IMS and VSAM, object-oriented
databases such as POET, or XML databases (although MetaTrieve will process
XML data types stored in a relational database as tables and columns).
Yes, you can search up to four distinct databases with one search
execution. For example, with one search, you can scan your company’s
mainframe DB2 database, data warehouse SQL Server database, eCommerce
Oracle database and a shared-drive ACCESS database. Once your
database are setup in ODBC and on the administration page, you can simply
click on all of these on the search page to scan your enterprise.
You simply click on a search result row and this takes you to
the full table display with a template query which you can modify.
No, you can go directly to a query by selecting the “Query”
menu option, or by clicking on “Search” without entering a
search term.
The response time varies depending on the desktop processor and
memory, the size and type of the company databases, database access contention,
and if synonym matching is desired. For example, on a Dell P4 with
256 RAM a search against a typical CRM or data warehouse SQL Server database
with over 2000 business columns takes about 1-3 seconds; against a large
mainframe operational DB2 database with over 30,000 business columns,
the response time may run 4-10 seconds.
Generally, large databases have many schemas/owners, only one
or two of which actually contain the production business information.
Your database administrator may provide you with this schema name and
you can select it on the administration page. This will speed up
the search as will not selecting synonym matching (on the admin page)
unless needed.
While MetaTrieve calculates a confidence value for each candidate
match, only the human eye can truly discern the meaning of a metadata
term. Often, you will need to display the entire table (by clicking
on the match result row) or query the data to get a complete picture of
the context and meaning of a database element.
Often, the search itself will display data that is related semantically
to the business item(s) you are seeking. However, MetaTrieve also
displays explicitly related tables and columns Foreign Keys) that have
been identified as such by the database designer. You can view these
related elements via “Related Data” button on the full table/query
page. Once on the “Related Data” page, you can continue
to follow a chain of relation by selecting a desired row (this will bring
up the related table which, in turn, may have other related data.
Yes, you can save the results sheet as test or as a spreadsheet
and open it later for further querying. You can also save the data
results of a query as text or Excel sheet or Microsoft Access table (if
displayed in Excel or Access).
You can perform any ANSI relational queries (including joins,
unions, sub-selects, updates, inserts, deletes, etc.) that you have database
permissions for. For specific vendor databases, you can also perform
those unique SQL statements.
Yes, you can save SQL into a folder and open it later for querying.
You can also create generic queries by entering “?” prompts
and MetaTrieve will request you to replace these prompts at run time.
You can see the results of your query in a MetaTrieve window,
in Microsoft Excel or Microsoft Access. If you expect the results
to return a large volume of data, Access will hold the largest result
set. MetaTrieve has a 32,000 row limit, Excel 75,000 and Access
is effectively unlimited.
You can view multiple tables by clicking on multiple search result
rows or by selecting different tables on the “Select Table”
window (from the “Query” menu option) or by selecting tables
from “Related Data” display rows.
If you wish to create a Microsoft Access copy of a database table
or joined table in Access, first you need to create a query on the MetaTrieve
query page and display the results in Access. This will display
as a pass-through query where you can convert it to a table. Follow
the Access instructions for converting a pass-through query to a table
as outlined in the Help documentation for Access (type in “pass-through
query” in the indexsearch).
|
 |
 |