India
17h
3238
Why is it so G*damn difficult to move money out of India
Health & Wellness
16h
952
Issues with sleep
2024 Presidential Election
18h
1363
Uh oh: President Trump leads Biden 49% to 43% in a two-way race.
AMA
Yesterday
3331
I have worked at TikTok US core tech for 3 years. AMA.
Tech Industry
11h
532
How’s capitalism going?
I have a usecase to support full text search in 150 million records. I’m constrained to use relational database. Elastic search seems like a right fit here , but thats not an option for me. Folks with relational database experience, Do you think this is a feasible in database like postgres / Oracle? Thanks!
Bump
Postgres supports full-text search. Look at ‘tsvector’.
It’s very much possible in rdbms but that’s not for what they are built for. Oracle supports text search but terrible in scaling. The usual transaction processing power becomes a liability here. It’s called text or context option I think.
150mil records is nothing for RDBMS to handle. I believe most of them have it. What about scalability requirements and latency for indexing ? Many years ago while building system using SQL Server we could not use its full text search as it was not transactional: index was updated with like 1h delay or so.
This is to support filtering and pagination for the UI. So the latency should be good, for better user experience or at least should not time out:) I’m assuming anything less than 2000ms should work.
It does not have to scale indefinitely, we would purge data if it goes beyond 150 million
Forced to use relational but the database is up to your discretion? Sounds like whoever is making the call doesn’t know any better. Just use elastic and tell them it’s relational.
👏🏻