interMedia Exchange Tool

Oracle has evolved its multimedia and unstructured data handling features considerably since the InterMedia days in Oracle 8i. Here’s how the functionality has transitioned into Oracle 19c, 23c, and 23ai:

1. 🔹Oracle InterMedia (Oracle 8i to 10g)

   • Originally designed to manage images, audio, video, and document content within the Oracle Database.
   • Provided content extraction, search, and management APIs for multimedia data types.
   • Deprecated in Oracle 11g, eventually removed in later versions.

2. 🔹 Oracle Text (Rebranding and Evolution)

   • Current equivalent for unstructured text data such as documents, HTML, PDFs, etc.
   • Fully supported in Oracle 19c, 23c, and 23ai.
   • Features include:
     • Full-text indexing
     • Fuzzy searches, stemming, thesaurus support
     • Document section indexing
     • Integration with LOBs, XML, and JSON.
   • ✅ Oracle Text is the direct and supported evolution of InterMedia's text capabilities.

3. 🔹 Oracle Multimedia (Removed as of Oracle 19c)

   • Replaced InterMedia for multimedia data (images, audio, video) in Oracle 11g through 12c.
   • Deprecated in 18c, removed in Oracle 19c.
   • Oracle now recommends storing multimedia content outside the database (e.g., in object stores or filesystems) and managing metadata inside the database.

4. 🔹 Modern Alternatives in Oracle 19c/23c/23ai

   Legacy InterMedia Feature	Oracle 19c/23c/23ai Replacement
   Text indexing/search	✅ Oracle Text
   Multimedia management	❌ *No direct in-DB equivalent*
   Image/Video processing	🔁 *External tools + BLOBs*
   XML/HTML document parsing	✅ Oracle XML DB
   JSON document handling	✅ Oracle JSON / JSON Relational Duality Views (23c)

5. 🔹Oracle 23c / 23ai Specifics

   • Oracle 23c introduces Duality Views for JSON ↔ Relational mapping (useful for document-style content).
   • Oracle AI Vector Search (23ai) enables similarity search on unstructured embeddings, ideal for AI-powered document discovery, not present in InterMedia.
   • Oracle Autonomous JSON Database (AJD) is available for document-centric workloads (think of modern NoSQL-style unstructured data processing).

✅ Summary

Capability	InterMedia	Oracle 19c/23c/23ai
Text indexing	✅	✅ Oracle Text
Multimedia in-database	✅	❌ Removed (store externally)
JSON / XML	❌	✅ Oracle JSON / Oracle XML DB
AI & semantic search	❌	✅ Oracle AI Vector Search (23ai)
Content-based retrieval	✅	✅ Text-based only (no image/audio)

If you are migrating from InterMedia, the best strategy is to:

• Move media assets to external storage (e.g., Oracle Cloud Object Store).
  • Use metadata tables with `BLOB` references or URLs.
• Use Oracle Text for textual content extraction/search.
  • Leverage JSON/XML DB for structure and indexing.
  • Optionally incorporate Oracle AI Vector Search for advanced similarity matching.

✅ Terminology Shift:

Oracle Version	Nomenclature and Description
Oracle 8i	Oracle InterMedia: Suite for managing multimedia data (text, image, audio, video).
Oracle 9i–11g	InterMedia (deprecated gradually) → Oracle Text + Oracle Multimedia: Text indexing separated; multimedia handled via Oracle Multimedia.
Oracle 12c–18c	Oracle Text + Oracle Multimedia (deprecated in 18c): Multimedia was removed; Oracle Text retained and enhanced.
Oracle 19c–23ai	Oracle Text: Sole supported feature set from InterMedia, focused on text indexing, search, and document handling.

🟢 Oracle 23ai Status:

• ✅ Oracle Text is the official and supported replacement for InterMedia’s unstructured data capabilities.
• ❌ Multimedia storage (audio, video, image) is no longer handled natively in the database—external storage is recommended.

In Oracle 23c and 23ai, Oracle Text integrates with JSON through new capabilities that allow full-text indexing and querying over JSON documents stored in relational or hybrid (duality view) structures.
🔍 How Oracle Text Integrates with JSON in 23c/23ai

1. Indexing JSON Columns
   Oracle Text can now create text indexes on columns that contain JSON data, enabling full-text search within JSON documents. This works even if the JSON is stored as:
   • CLOB
   • VARCHAR2
   • Native JSON data type (new in Oracle 21c+)
2. Syntax: Using CTXCAT or CONTEXT Indexes on JSON
   

   CREATE TABLE products (
     id NUMBER PRIMARY KEY,
     product_info JSON
   );
   
   -- Create a text index on the JSON column
   CREATE INDEX product_info_textidx ON products(product_info)
     INDEXTYPE IS CTXSYS.CONTEXT;
   

3. Querying with CONTAINS() on JSON
   You can perform fuzzy, wildcard, or boolean text searches directly:

   SELECT id
   FROM products
   WHERE CONTAINS(product_info, 'wireless AND mouse') > 0;
   

This lets you search across deeply nested keys and values inside JSON.
🧠 Oracle 23ai Enhancements

• ✅ JSON Relational Duality Views

  • JSON and relational models are synchronized via duality views.
  • You can apply Oracle Text indexing to both the underlying JSON columns and duality view virtual columns.

• ✅ Oracle AI Vector Search + Oracle Text

  • Combine Oracle Text with Vector Search in 23ai to perform semantic search across unstructured text and JSON descriptions.
  • Example: Combine CONTAINS() for keyword match and vector similarity for meaning-based match.

🛠 Example: Full Workflow

CREATE TABLE documents (
  doc_id NUMBER PRIMARY KEY,
  content JSON
);

-- Add text index
CREATE INDEX content_text_idx ON documents(content)
  INDEXTYPE IS CTXSYS.CONTEXT;

-- Sample Query
SELECT doc_id
FROM documents
WHERE CONTAINS(content, 'deep learning') > 0;

If you use a duality view:

CREATE OR REPLACE JSON RELATIONAL DUALITY VIEW doc_view AS
  SELECT d.doc_id, d.content FROM documents d;

-- Now you can search via `doc_view` as well, assuming indexing is inherited.
🔐 Important Notes

• JSON data must be well-formed.
• You can use section groups to target specific keys inside JSON.
• Consider indexing strategies for large JSON blobs for performance.

When you run a Web site, you coordinate files that contain video, audio, and graphic data. Loading and retrieving these kinds of data from the database used to be difficult and required special programming skills. With Oracle's interMedia tool, a great deal of that specialized programming has been provided for you. Now you can use familiar PL/SQL calls to stored packages to modify or query audio, video, and graphic data formats. The interMedia toolset includes several related components, as you can see in the MouseOver diagram below.

The interMedia tool allows you to retrieve data from just about anywhere and load it into the database. In particular, the Java client side of interMedia makes it easy for you to load images or other media data from your client's desktop into your database. The table shows some of the useful features of inter media.

Feature	Description
Query attributes of image, video, or audio type data	interMedia provides methods for the object types ORDAudio, ORDVideo, and ORDImage. These methods allow you to attach comments, descriptions, height, width, length, compression type, format, and other important information about each image, video, or audio record. These attributes can be queried by using methods as well.
Get data from the database, modify it on the desktop, and replace the modified data in the database by using a Java application	The Java client for interMedia builds a copy of the media file on the desktop. If the user can update the copy, the Java client locks the database record so no one else can modify it until it has been returned.
Extend interMedia for all types of data	You can write customized subclasses and methods for the interMedia ORDMultiMedia class so that the database handles any kind of data you have.

The Visual Information Retrieval tool is related to the interMedia tool. Both can store and retrieve image files in the database using objects and methods. The Visual Information Retrieval tool has additional query capabilities based on image content.

inter media Terms

In your travels around the internet, you have probably used many powerful search engines. Many of today's top search engines use the features such as those in interMedia text to provide the reliability of results that is expected from them. A case study that builds a search engine will be shown in a case study later in the module. If you were a computer book addict, and were looking for info, you could search for the phrase books. The problem is you will receive results on everything from cooking books to books on trains in no useful order.
By utilizing interMedia text, we can focus that search into a more useful result set. If you want to narrow down the search, you would specify a search like

books + potter

programming. This tells the search engine that you are searching about books, and want Potter books ranked highest, also you want to eliminate or lower the rank of results with programming books in the search.
Modifiers, such as + and , appear to be very useful in a search, but we must keep in mind that most people will not know that these special modifiers exist, and the average person will not want to spend time on reading through documentation or help.
Users will not usually spend the time to learn the ins and outs of proper usage. The average user will want to use natural language searches for 'themes' and 'gists'.
A natural language search is a search in the form of a question:

Where do I find out about books?

The theme of the previous question is 'books'. Oracle interMedia has integrated the ability to add to indexes themes of documents or concepts within a document, though it is up to you to plan the design around your intended user.
The next lesson tells about the Internet File System (iFS).