Hello, I have some basic questions for the group. I would appreciate any advice you can give me.
We have an Oracle RAC database that has a number of schemas on it. Various "things" query the structured data stored in these schemas, 10s of thousands of times per day. Two of these schemas in particular we wish to replicate into a Lucene-based enterprise search environment. This in total will cover about 750 million "records" that span a few dozen tables in each schema. As I said, we have structured data here. No documents, no BLOBS and no CLOBS. We have front-end search tools already in place that provide the user with filtering options that, when submitted by the user, run carefully optimized (and indexed) dynamically-built queries against the data in these schemas. These are on all the common fields you'd imagine when searching for user data--name, etc. As I said, we would like to replicate this data into a Lucene-based environment and revise the front-end that formerly was hitting Oracle to instead call RESTful APIs that we develop, to go against this new Lucene-based tier. We want to retain all of the previous faceted search abilities, and not have anything run more slowly than before. In addition to retaining all of the existing filtering options for the user, we would like to introduce an additional single google-style search window that really takes advantage of the fancy new indexes and features that SOLR can offer. At this point I am trying to understand how the indexes work for structured data in Lucene, and whether this is feasible to basically take all search processing and querying out of Oracle. Most of the Lucene/SOLR documentation that I've found puts emphasis on all the great things it can do with unstructured data, and does not address what it can do with structured data. Can anyone please give me some feedback or point me in the right direction? Second question-- members of our group are trying to sway us to using Coveo Enterprise Search, another Lucene-based tool, rather than SOLR. I realize this is a SOLR forum, and I was hoping someone might be able to give some insight as to why SOLR (or Coveo) would be the better option. We will be using custom written interfaces, so SOLR's out of the box search interfaces would not be helpful to us. Thanks for your time, Eric
