Re: PL/SQL general question
Date: 1 Oct 2001 03:59:36 -0700
Message-ID: <9p9iao02amr_at_drn.newsguy.com>
In article <3BB7C9D0.7C0B0159_at_sqlman.com>, Tom says...
>
>Sorry, I must not be a serious developer -- but what is interMedia Text?
>
interMedia text is a part of the database that supports the efficient indexing of large amounts of textual information stored in a single column.
It provides the keyword extraction you are looking for. It uses a bitmapped index instead of the inverted list it sounds like you were going to use (bitmaps are smaller, faster). It supports all kinds of searches in SQL against this text then. You can set it up to index only special words, or all words (minus stop words).
see
http://technet.oracle.com/doc/oracle8i_816/inter.816/index.htm
for more info
>Rich Foster wrote:
>
>> This sounds like a homework assignment. Any serious developer would
>> immediately consider interMedia Text as a solution to index and do keyword
>> searches.
>>
>> Rich
>>
>> mustang4u_at_home.com wrote:
>>
>> > I am new to programming and have been asked to write a program to
>> > produce keywords from several description fields in a database. The
>> > main requirement is speed! My question is weather PL/SQL is an
>> > appropriate language to do this type of manipulation. Or if it is
>> > better left to a higher language like Pro c. From what I have read
>> > this is not impossible to do in SP/SQL but my impression so far is the
>> > language was meant to handle more database report type programs.
>> >
>> > The program would have to :
>> > Read strings from a table
>> > Tokenize them
>> > Remove any duplicates
>> > Strip special characters
>> > Write them to a table with a unique identifier
>> >
>> > Any opinions would be great
>> >
>> > Thanks Kevin
>
-- Thomas Kyte (tkyte_at_us.oracle.com) http://asktom.oracle.com/ Expert one on one Oracle, programming techniques and solutions for Oracle. http://www.amazon.com/exec/obidos/ASIN/1861004826/ Opinions are mine and do not necessarily reflect those of Oracle CorpReceived on Mon Oct 01 2001 - 12:59:36 CEST