Omnidex Text

Overview

Omnidex provides several powerful text searching capabilities that allow developers to provide users with very fast and flexible text search applications. The Omnidex text search features include keyword, synonym, proximity, phonetic, misspelling and spell-checked searches, as well as useful thesaurus and form search capabilities and a dictionary for definition retrievals. Proximity searches also enable rank and relevancy score retrieval, as well as context retrieval with optional an highlighting feature.

All of these text search capabilities are available on all textual data type fields, including extremely large amounts of data contained in CLOB data type fields as well as text documents external to the database. The text search capabilities are available for use in any Omnidex application, some by default and some through specific Omnidex extended functions.

At this time, Omnidex supports plain text and html text documents external to the database.

Summary of Features

• Omnidex text can index text in any textual column in a database, as well as in text documents external to the database, up to 64MB.
• Keyword searches can be performed against any of these indexed columns if the ;kw option is installed.
• Editable stop-word lists can be used to prevent "noise" words like "and" and "the" from being indexed in keyword indexed columns.
• Searches will be case-insensitive if either the ;kw or ;ci option is installed.
• Proximity searches can be performed on columns installed with the ;px option.
• Rank and relevancy scores can be retrieved when criteria is applied to a column installed with proximity, using $SCORE.
• Formatted excerpts showing the context of qualified words or phrases can be retrieved when criteria is applied to a column installed with proximity, using $CONTEXT.
• Phrase searches can be performed on any column installed with the ;px option.
• Synonym searches can be performed against any column installed with an MDK index, using the built-in thesaurus or managed synonym lists.
• Misspelling searches can be performed against any column installed with an MDK index, after the ODXSQL command UPDATE TEXT has been executed against the database.
• Spell-checked searches can be performed against any column installed with an MDK index.
• Form searches can be performed against any column installed with an MDK index.
• Phonetic searches can be performed against any column installed with an MDK index.
• Metadata can be retrieved using SQL, for synonym, form, spelling, stopword and definition searches.

Limitations

All columns in a table, except for the last column, must have an offset of 65,535 bytes or less. This means that there may be only one data type with a length greater than 65,535 AND it must be the last column in the table definition. Therefore, there can be only one column per table, the last column, of any of the following data types: CSTRING with a length of 32,768 bytes or more, CLOB, OMNIDEX CLOB, BLOB, or OMNIDEX BLOB.

The size limit depends on the data type See Supported Data types for details.

Top