You can use varchar types and inside your saveDoc() function you can save the document in smaller chunks in multiple rows with an ORDERING column. Or call it PAGE column. When you readDoc() you can assemble them back again. Table data will look like:
DOCID TEXT PAGE
1 "Hello " 1
1 "World." 2
2 "Another " 1
2 "Another " 1
2 "Doc." 2
When you search for "text" you will get back DOCID and PAGE which can help you locate relevant chunk quickly. You can maintain other information on a document identified by DOCID in another table.
Regards,
Ali
Mag Gam <[EMAIL PROTECTED]> wrote:
Wolfang:
Hmm! Very interesting. What about if I use something like Lucene to index my data?
I am trying to build a simple database to put my "notes" and do a quick search on them when I need some info....
ANy ideas now???
On 9/25/05, [EMAIL PROTECTED] <[EMAIL PROTECTED] > wrote:Hi Mag,
As long as the doc ( http://db.apache.org/derby/docs/10.1/ref/rrefjdbc96386.html)
says true,
you cannot index the CLOB column.
---here is a snippet of the doc-------------------------
* Restrictions on BLOB, CLOB, (LOB-types):
LOB-types cannot be compared for equality(=) and non-equality(!=, <>.
LOB-typed values are not order-able, so <, <=, >, >= tests are not
supported.
LOB-types cannot be used in indices or as primary key columns.
DISTINCT, GROUP BY, ORDER BY clauses are also prohibited on LOB-types.
LOB-types cannot be involved in implicit casting as other base-types.
--------------------------------------------------------
Regards,
Wolfgang.
__________________________________
F or All Sports Fans!
http://pr.mail.yahoo.co.jp/yells/
