Cataloging Home

Tools for Catalogers

Unicorn search strategies : Workflows keyword searching

Much of the power and advantage of Sirsi's Unicorn is in its keyword searching functionality, powered by the site's ability to customize indexing through format-specific indexing policy tables. These policy tables enable one to pinpoint certain types of information provided in both the bibliographic or title record and its corresponding call number and item records.

Architecture of Unicorn bibliographic information

Before we can understand how searching works in Workflows it is important to understand the relationship of the three levels of bibliographic description associated with a book or other information resource and what types of unique data they contain. In Unicornspeak, these are called the title, call number, and copy records. The distinction between the latter two is a Unicorn peculiarity.

Title record

The title record consists of the MARC bibliographic record describing an information resource, along with management information appropriate at this level. The other data includes several fields that Unicorn uses internally to display holdings information in the OPAC plus date stamps indicating when activities, like record creation or editing, have transpired. The added MARC fields are:

The following table shows management information which is primarily used in reports for restricting results to records added by a particular group, in a particular date range, or by format. The last two columns indicate whether the field is accessible through a keyword search or report.

Call number/volume record

The call number record is associated with an owning LIBRARY and contains that library's call number and volume information. There are only four fields at this level:

Copy record

The copy or item info contains attributes associated with a particular copy of a volume, including the item ID or barcode. This is the source for several of the filters used in both the catalog and reports, including the "Limit by Material" and "Reference Resources Only" filters in the public catalog.

General search

These are the most powerful of all Unicorn searches, as these search all indexed MARC fields. A table of keyword searchable fields is available on this site. We were fairly liberal in including fields for keyword indexing, so in addition to the usual author, title, series, and subject fields, we have access to all control number fields (010, 020, 022, 035), many of the coded fields (043, 045, 048, etc.), all call number fields (050, 055, 060, 070, 082, 090), most note (5xx) fields, edition (250), imprint (260) and pagination (300), serial frequency (310), and even the URI field (856). An unqualified search will retrieve information from any of these fields, but Unicorn also allows for searching specific MARC fields using braces or "squiggly brackets". Note that more than one field can be addressed at a time using this methodology and the use of spaces before the bracket and between field tags is optional.

• ncdocs {086} and forestry {650,651} (all NC state documents have the qualifier "ncdocs" in the document number field)
• burford{505} (retrieves from author statements in table of contents; useful for conference proceedings)
• oxford university press{260} and shakespeare
• IEEE {111, 711} and GaAs
• chiefly ill {300} and architecture {650} (retrieves architecture books that are heavily illustrated)
• brando {511} (retrieves actors names from movie performer note)
• netlibrary{856} (searches for a string within the 856, in the |uURL or |y|z|3notes)

Logical operators

Boolean operators

AND retrieves records containing all of the specified terms and in one or more MARC fields.

OR retrieves records matching any or all of the specified terms.

NOT retrieves records containing the first search term but not the second.

XOR (exclusive OR) retrieves records matching any of the specified terms but not all of the specified terms.

Positional operators

SAME is the unstated default operator in Unicorn and will retrieve records in which a single MARC field contains all of the specified terms. If the field contains multiple sentences, the search will continue through all. If you need to search across multiple MARC record fields, use AND.

WITH retrieves records in which a single MARC field contains a sentence with all of the specified terms.

NEAR retrieves records in which a field contains all of the search terms next to each other, though not necessarily in the order in which you entered them.

ADJ retrieves records in which a field contains all of the search terms adjacent to each other and in the order they were entered, even where the words may not be in the same sentence or subfield.

Relational operators

(>, <, =, <>, >=, <=).

• smart adj materials
• >1990{260}
• (shakespeare and lear {245}) not (videorecording or penguin)

Substitution (wildcards) and truncation

The substitution character (?) enables the user to replace a single missing character in a search:

• monomolecul$ (retrieves "monomolecule", "monomolecules", "monomolecular", etc.)
• anders?n (retrieves "Andersen" and "Anderson")
• wom?n (retrieves "woman" and "women")
• 196? (retrieves "1960", "1961", "1962", etc.

The truncation symbol ($) may be used to replace any number of alphabetic characters, either within a term or at its end. Unlike the ?, $ does not work for truncating numbers. It can be used for searching word roots, singular/plural terms, variant U.S./British spellings, misspellings, or base Dewey or LC call numbers. Optionally, use of a number after the $ enables the user to specify the number of characters you intend to replace. Here are some examples:

• command$ (retrieves "command", "commands", "commandment", "Commander", "Commandante", etc.)
• substitut$3 (retrieves "substitute", "substitutes", "substitution")
• brit$n (retrieves "Britain", "Britian", "Briton", "Britten", "britischen")
• col$2r (retrieves "color", "colour", "collar")
• 541$ {082} (retrieves "541.3451", "541.22", etc. from the Dewey class number field)
• QA76$ (retrieves "QA76", "QA76.5", "QA76.76.H45", etc.)

To perform a search for a term with a dollar sign or question mark in it (e.g., $1,000,000), enclose the search term in double quotes.

Searches without search terms

Unicorn is one of the few integrated library systems that allows users to perform searches using filters only, or even searches without filters at all. These are extremely useful in enabling staff to quickly count titles by library, format, circ type, year of publication, or by any combination of the above. Filtered searches can be done using any of the categories under "Set Options for Item Lookup", the binocular icon found at the upper left of the any of the search windows in Workflows. You must have first selected the "Search" option to use all the filters. Your choices for browse are limited.

To get a count of or to see all titles held at a branch or library without using filters, you will need to use "#0" (pound-zero) as your search argument. These searches take longer than filtered searches, so be prepared to wait for up to a minute for your results. Note, that if you are using any filter from the "Set options" icon, the "#0" is not needed and will only slow down your search. To get a count of all titles in the catalog simply leave the default of ALL under library and enter "#0" in the search box. If it's volumes or copies you wanted to count, sorry, you will need to run a report! Note that a #0 count will include ALL title records associated with a library, including Reserves, titles on order, withdrawn or lost, unused Marcive Shipping List records, and brief on-the-fly records that may have a corresponding bib record elsewhere in the catalog.