Search Term Syntax
A natural language search string can be made up of a single word, a list of words, or a natural language question. For example, entering the search term:
only returns documents that contain the word "computer." However, entering the search terms:
how do i attach a printer to my computer
returns documents about printers and computers, but ranks documents which discuss attaching printers to computers at the beginning of the result list.
Phrase searches are used to ensure that one word is followed by another, with no other words in between, in the returned document. You conduct a phrase search by enclosing search terms in double quotations. For example, entering the search string:
only returns documents that contain the phrase "cordless telephone." Documents that contain the words "cordless" and "telephone" separately will not be returned. An alternative method of achieving this same result is to use the adjacency operator ADJ between terms. For example,
cordless ADJ telephone
will only return documents that contain the phrase "cordless telephone." Documents that contain the phrase "telephone cordless," or the words "cordless" and "telephone" separately will not be returned.
Note: The adjacency operator I>ADJ must be entered in UPPERCASE letters, otherwise it will be treated as simply part of the search phrase.
For data collection documents that are structured in a semi-regular format, regular portions of the document may be tagged as fields. You can use these fields to restrict a search to only those documents containing a specified value in a particular field. In the WAIS search syntax, performing a restricted search based on the value of a field or set of fields is called a fielded search. In the search server, each field has a corresponding searchable column or zone in the base table of the database. A fielded search in the WAIS search syntax is represented as:
fieldname = fieldvalue
where fieldname is the name of a searchable column or zone in the base table, and fieldvalue is the value to be searched.
For example, electronic mail databases contain "to" and "from" fields. A fielded search of the "to" field for a value of "Martin" is given by "to = martin".
Right truncation is specified by ending a word with an asterisk (*) as a wild card character. This is interpreted as a search on words matching the base characters before the asterisk, and ignoring any trailing characters.
For example, a search string containing the phrase "geo*" may retrieve documents containing the words geographer, geography, geologist, geometry, or geometrical.
Left truncation is specified by beginning a word with an asterisk (*) as a wild card character. This is interpreted as a search on words matching the base characters after the asterisk, and ignoring any leading characters.
For example, a search string containing the phrase "*fix" may retrieve documents containing the words prefix, suffix, or postfix.
Mid truncation is specified by placing an asterisk (*) in the middle of a word as a wild card character. This is interpreted as a search on words matching the characters before and after the asterisk.
For example, a search string containing the phrase "com*ent" may retrieve documents containing the words commitment, competent, or compliment.
The Mathematical operators >, <, =, <= and >= aid in establishing mathematical relationships between concepts expressed in natural language.
The Boolean operators AND, OR, and NOT aid in establishing logical relationships between concepts expressed in natural language. These operators are especially useful in narrowing or expanding a search.
Note: Boolean operators must be entered in UPPERCASE letters, otherwise they will be treated as natural language queries.
The AND operator is useful in restricting a search when a particular pair of phrases are known. For instance, when searching for documents containing information on the weather in Boston, a search string such as "weather AND Boston" only finds those documents containing both the words "weather" and "Boston". An alternative syntax for this search is "weather && Boston".
The OR operator is used to join two different phrases of a Boolean search together. A search string such as "hurricane OR tornado" finds all documents containing either the words "hurricane" or "tornado" (or both). An alternative syntax for this search is "hurricane || tornado".
The NOT operator is used to reject documents that contain certain words.
The search string "basketball NOT college" finds all documents that contain
the word "basketball," but do not contain the word "college."
The operator TO is used to establish a range of values to search for. The range is inclusive of the two values specified.
Note: The range operator must be entered in UPPERCASE letters, otherwise it will be treated as natural language search string.
The search term syntax used to specify a date or numeric range is as follows:
FieldName ComparisonOperator Value
where the ComparisonOperator is one of:
and Value is a numeric value or a date (see Date Formats, below)
Also supported is the TO operator:
FieldName = value1 TO value2
which is equivalent to the following:
( FieldName >= value1
) AND ( FieldName <= value2 )
Date Format Syntax
If Value is a date, the syntax is specified as follows:
year-month-day day-month-year month-day-year
where the month and day contain 1 or 2 integer characters, and the year
contains 1 to 4 integer characters. Possible separators are '-', '.' and
From the context, SurfServer determines which format is being used.
For example, the following examples show three different unambiguous representations
for the 20th of February, 1995.
95-02-20 is interpreted as year, month, day.
20.2.1995 is interpreted as day, month, year.
02/20/1995 is interpreted as month, day, year.
Precedence of Operators
When search terms contain more than one operator, operator order (precedence) may be controlled with parentheses.
A Canned Search is a URL (Internet address) that is returned as a result of running a search operation. The URL contains all the search parameters needed to re-run the search. Users can save the URL as a bookmark and use it later to access search results.