I already have a product in production where i used Entity Framework with SQL Server as db. I've used full text search wherein i store binary data (filestream) along with the file extension which is used by the Full Text Search Engine to index. Not to mention that was a cake walk.
Now, am planning to move to mysql (cuz of obvious reasons -- cost, open-source etc). The product is in SAAS model (although this is the time i can actually take definite call as traffic isn't high) so volume will be high, so search engine should be scalable.
Migration to mysql is easy (will be using INNODB, again for obvious reasons), am only stuck with Full Text Search as right now only the binary data is stored in Sql Server. Although INNODB supports full text search in version 5.6 onwards I did not find exact way to Full Text Index (that is using binary).
Am not sure of using third party full text search engine (Lucene, Sphinx etc) as my searches will be combination of structured and unstructured, for ex: All the Customers from Ohio (structured data as i capture and store the information in RDB) and who have "insurance" in their set of documents (unstrucutred) uploaded. In SQL Server i use "ContainsTable" which gives me ranked results.
I have following questions :-
- Will this movement really be fruitful in the long run? Lets say i do migrate the existing data (in 4 figures).
- Although INNODB provides all the necessary features, does that match in SQL Server ease of use, administration and scale?
Some more questions...
- Can i move the full text table to mysql as is? I guess i cant index blob columns.
- Should i use mysql full text or any third party engine?