A few suggestions for searching LONG VARCHAR fields.

1. I suggest that LONG VARCHAR fields be searchable if and only if they are
indexed. The user/programmer selects which LONG fields are indexed and how
many characters are included in the index. MySQL does this:
http://www.mysql.com/doc/C/R/CREATE_INDEX.html
[ CREATE INDEX part_of_name ON customer (name(10)); ]
This is one option that would work for many people. 

2. I think a better solution would be to incorporate a new module like an
Informix Excalibur text blade
(http://www-3.ibm.com/software/data/informix/blades/excaliburtext/),
Thunderstone's Texis (http://www.thunderstone.com/), or the open source
HTDig (http://www.htdig.org/) that is a real text search engine with search
syntax like: proximity, in, not in, in same sentence, in same paragraph,
fuzzy spelling, etc. Of course this would be a major effort but that would
make SAPDB a solid contender for storing unstructured and semi-structured
content. I have heard that 80-90% of corporate data is not structured
(meaning it can not fit into columns and rows) like email and office
documents. Nearly all database vendors have a solution or are actively
developing a solution for searching large text objects. SAPDB should move in
this direction as well. With MS Office documents moving toward an XML
format, searching them in this context should be easier in the future. (Of
course you may need to get a patent license to parse their XML.)

Dan


_______________________________________________
sapdb.general mailing list
[EMAIL PROTECTED]
http://listserv.sap.com/mailman/listinfo/sapdb.general

Reply via email to