One of our two databases uses context indexes on a column.
If your not familiar, this is part of the Oracle interMedia option. The idea
here is to be able to search for words within a text block. The text block
is in a row, in one of the columns. It can also be HTML, a Word document,
etc. When you index the column, it speeds up finding matching words.
Once that's done, you can search like this:
SELECT SCORE(1) title FROM news_items
WHERE CONTAINS(text, 'Bolivia', 1) > 0;
This would find all rows in the table "news_item" that has the text
"Bolivia" in the "title" column.
SELECT SCORE(1) title FROM news_items
WHERE CONTAINS(text, 'Bolivia OR Peru', 1) > 0;
This would find all rows in the table "news_item" that has the text
"Bolivia" or "Peru" in the "title" column in any order.
This is the most basic usage of the interMedia (you can get it to do a
soundex to match a sound, or get a "near" match or fuzzy match rather than
an exact match, etc.)
But I'm looking for the basic "return all rows where the text in this column
returns this word or these words".
My two thoughts are, write my own function in PERL, or store the text in a
file outside the database and use some OS-features to do the search
Any suggestions?
David
