Administration Guide : Xinet Database Overview : About the database and Search logic

About the database and Search logic
Xinet offers two search methodologies:
With Solr Search enabled, Portal users have faster and more comprehensive searching using the Quick Search functionality. The Solr Search index is generated using an n-gram tokenizer which segments words into 3 - 18 characters, therefore producing a more robust index. Solr Search provides structured metadata filtering available in the Portal Filters tab and Modified on and Created on date selection buttons located in the Portal browser view. By default, the Solr search AND operator is used. In addition, Solr Search provides count values for metadata, grouped by type, found for a query. This functionality is commonly known as Facets. These counts are displayed in the Portal Filters tab. When Solr Search is enabled, Portal users can also use the Advanced Search functionality which uses the FullText indexes, to complete complex search queries.
With FullText indexes enabled, Portal and Pilot users can conduct advanced searches using logical constraints.
About Solr Search indexes
The Solr Search engine uses an n-gram tokenizer to generate a complex index. Words associated with an asset, such as metadata, filenames, content, annotations, and comments are indexed into 3-18 character tokens. As a result, Portal users completing a Quick Search are presented with results of assets found using a combination of the characters specified in their search query.
Depending on the number of facets established for Portal users, the database will increase in size. For more information on how to mark particular Data Fields as Searchable in its Permission Set, see Customizing Templates for Individual Users or Groups — the Permission Sets subtab for details.)
To learn more about the Solr Search engine, see
About FullText indexes
Enabling the database and FullText Indexes provides faster, better-targeted searching facilities for Xinet clients. Enabling the database is essential for searching in Portal. Without the database, Portal users can only browse and Xinet users can only conduct searches using Filename (but, without logical constraints), Files or Folders, Filetype, Search In (Online vs. Archives), Date, and Comments. With the database enabled Xinet users can also conduct searches using Description, Filename (with logical constraints), Search All Keywords, various nativeadmin-determined Keyword or Data Field searches and some file system attributes such as Created on, Modified on, and Accessed on. Each user may, depending on several factors, also be able to further constrain any search option. With FullText Indexes in use, he or she can generally anticipate faster results, but in cases where their use is not helpful, the search engine will not use them.
With the database enabled on volumes Portal users have a greatly-extended set of targets for searching with more flexible search logic and syntax. Among other things, they can search through extracted text from InDesign, PowerPoint, Microsoft Office, PDF and HTML documents, for file and datafield events, for video/interactive file information and events, for Annotations, for information about Approvals, for details about color space, and for information about Archive activities. In addition, they can save their search criteria and search results will be cached so they can be sorted and paged through.
For more information:
Searches are influenced by:
Whether a FullText Index of the database exists (See Native Search for more information about FullText searching.)
The kind of Search target established with the pop-up list by the user.
Whether a particular Data Field has been marked Searchable in its Permission Set. (See Customizing Templates for Individual Users or Groups — the Permission Sets subtabfor details.)
The type of Data Field; Type: introduces the various types.
Once a user selects a Search target (Filename, Files or Folders, Filetype, Search In, Date, Comments, etc.) he or she may then apply appropriate Search constraints before conducting the Search. Some constraints, such as searching for Files vs. Folders, or Online Files vs. Archives are self explanatory; however since some of the logical constraints may not seem so straight-forward. For information on logical constraints see “Constraining Searches” in the Xinet Client Guide.
The parameters a user can search, the logic he or she can use within the query and the speed with which Xinet returns results depend upon whether or not FullText Indexes exist. Even when FullText Indexes exists, there may be cases where it does not help, and the search engine will not use it then. Please note that Xinet searches can only use one FullText Index per search, in which case the Index must cover all the text fields in that particular usage. Xinet Portal, however, can make use of multiple FullText Indexes.
Please keep in mind:
Regexp searches may be slow, depending on the construction of the query, both with and without FullText Indexes.
When the user employs the Archived filter during an Advanced Search, a separate group of pop-up Search Logic options appear beneath other Advanced Search options. Unlike other selections selected during an Advanced Search, these will apply to the entire Search, regardless of the ANDs and ORs used.
Contains searches will be quite slow without a FullText Index.
Ends With will be quite slow without a FullText Index.
Is NULL does not use FullText Indexes.