Table of Contents
The search engine for help.unc.edu is powered by [ http://lucene.apache.org/ ] Apache Lucene . While a simple keyword search is often enough to locate the right document, sometimes a slightly more complex query is required. Lucene provides a number of advanced query features beyond the simple keyword search. This document will highlight the basics of using the advanced query syntax.
Note
In May 2006, the search engine for help.unc.edu was replaced. The old advanced search page, no longer being relevant, has been removed. Advanced searches can now be performed in the basic search box by using the advanced query syntax covered in this document.
Lucene defines a
term
as a single word. Entering multiple terms in the search box will retrieve documents which contain at least one of the terms. For example,
oracle calendar
will search for documents with
either
oracle or
calendar . A
phrase
, however, consists of multiple terms surrounded by quotes. For a document to match, it must contain the phrase, exactly as entered in the search box. So, in contrast to the previous example,
"oracle calendar"
(note the quotation marks) will only return those documents which contain the phrase
oracle calendar .
Unless otherwise specified, terms and phrases are searched for throughout the entire document. Lucene supports field specific searches if you need to narrow your search down to a specific part of the document, such as the document's title. To search for a term or phrase in a specific field, enter the field name, a colon and your search term or phrase. For example,
title:"windows xp"
will only match documents with the phrase
windows xp appearing in the title. The following fields are available for searching on help.unc.edu:
Table 1. Available Search Fields on help.unc.edu
| Field Name | Content Searched |
|---|---|
author
|
You can use this search to find all documents owned by a specific person. Try using a search phrase, i.e. put quotes around the name:
author:"Charles Lee"
|
author-uid
|
Search for documents owned by the designated Onyen. |
content
|
This will search only the text of the document, not the title or keywords. |
genre
|
The genre of the document. |
group
|
Search for documents owned by a specific documentation group, e.g.
group:"Knowledge Management"
|
id
|
Search for a specific document id. |
link-to
|
Each document on help.unc.edu has a Document ID, which can be found near the bottom of the document.
link-to will return a list of documents linking to the specified document. For example ,
link-to:5765
will find all documents which link to this document (#5765).
|
subject
|
The specific subjects attached to a document. |
tag
|
The specific tags attached to a document. |
title
|
This will search only the words appearing in the title of the document. |
Lucene recognizes the Boolean operators
AND ,
OR and
NOT . Please note, Boolean operators
must
be capitalized in order to be recognized as Boolean operators.
Tip
Note the difference between the following three queries, listed in order of increasing specificity:
-
computer labs -
computer AND labs -
"computer labs"
The wildcard characters
* and
? are also available for use in Lucene. The
* wildcard matches multiple characters and the
? wildcard matches a single character. For example,
network*
would match
network ,
networks or
networking , whereas the search term
network?
would only match
networks. (Why doesn't
network?
match
network , you might ask? The
? wildcard matches
only
one character, not zero characters. Perhaps a more useful application of the
? wildcard is for capturing variant spellings, e.g.
gr?y
.) Wildcard characters can be placed in the middle of words, too, but you cannot put a wildcard character at the beginning of a word.
Parenthesis can be used to group multiple search terms into logical groups. They are often a good idea if you are combining multiple terms with different Boolean operators, as the results will be easier to predict. For example, the search
photoshop AND (tutorial OR howto)
will return documents containing the word
photoshop and either
tutorial or
howto .
blackboard AND grades
would match documents containing both the words
blackboard and
grades anywhere in the document.
chang* AND onyen AND password
would match documents containing
onyen ,
password and any one of the variants of
change , such as
change ,
changes or
changing .
(title:"oracle calendar" AND title:client*) AND (mac* OR linux)
might be useful for finding Mac or Linux documentation on Oracle Calendar clients.
While this document covers most of the advanced search features you might want to use, there are a few more search options available which are documented on the [ http://help.unc.edu/?id= ] [ http://lucene.apache.org/java/docs/queryparsersyntax.html ] Apache Lucene site .


