Introduction to Omnidex Indexing

Index Creation

Omnidex takes a different approach to index creation than most databases. Most relational databases use the CREATE INDEX command to declare and populate the index. This means that each time an index is created, the table is sequentially fully scanned. While this approach means that the index is available as soon as the CREATE INDEX command completes, it becomes expensive when many indexes are being created.

Omnidex applications often have hundreds of indexes, and so Omnidex provides a more optimized, two-step approach to creating indexes. First, Omnidex indexes are declared using the CREATE TABLE or CREATE INDEX SQL statements. These statements declare all of the properties and options for the indexes, but they do not immediately populate the indexes. After the indexes are declared, they can all be populated at the same time using the UPDATE INDEXES command. This allows one table scan to be shared between all of the indexes, greatly improving performance.

Omnidex generally indexes at a rate of around 1 billion keywords per hour. A keyword is a word or value to be indexed. For non-textual fields, like state codes or postal codes, a keyword is a single value. For textual fields, keywords are parsed based on spaces and punctuation.

More >

Additional Resources

See also:

Back to top
admin/indexing/overview/creation.txt ยท Last modified: 2016/06/28 22:38 (external edit)