Another Oracle Text Tip

May 10, 2011

We ran a query about Oracle text search on Google the other day. We were surprised and almost delighted that Beyond Search appears near the top of the results list. We would prefer that the vendor appear at the top of the results list for information about the vendor’s own search engine. The goslings try, but the vendor has an obligation to provide a flow of on point, timely information about its products in our opinion. Believe it or not, a number of vendors are falling behind in providing information to their licensees. We hope this tardiness is not a trend.

Anyhoo, here’s another useful nugget of Oracle information for DBAs and users alike.  DbaSupport.com posted up an article titled “Oracle Text – Expanding Your String Searching Capabilities in Oracle Database”.

When the SQL WHERE clause falls short, the LIKE condition can generally be manipulated in creative ways to produce the results you need.  One mustn’t stop with LIKE; LIKEC, LIKE2 and LIKE 4 exist and are explained here.  The drawback is that there is a level of honing required to apply these, namely you need a pretty good idea of what you are looking for and where it is located.

So for broken and complex strings or searches in larger data sets, Oracle Text can help.  Directly from the article we absorbed:

“The four index types (generally domain indexes for that matter) and their query operators are:

  1. CONTEXT, using CONTAINS
  2. CTXCAT, using CATSEARCH
  3. CTXRULE, using MATCHES
  4. CTXPATH, using existsNode()

Some of the indexes also use parameters, and those will be examined in subsequent articles. The query operator names are somewhat intuitive in how they support the index type. In a context search (based on large coherent documents), we want to know if the text contains what we’re looking for. For many, but smaller documents (also of various types), we have to search through a catalog, so we perform a catalog search. In most contexts, we know that rules require a match, and one thing you can’t escape in XML is searching a path to see if something exists (at a node).”

We’ll spare you the needle/haystack reference.  Just consider tucking this tip away for later.

Sarah Rogers, May 10, 2011

Freebie unlike some of the engineering services provided by search vendors to licensees who are working to make these findability products “find”

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta