Query Syntax

Perceptive Enterprise Search Query uses a specific Query Syntax to create queries in one form or another. In a Menu Assisted Query the Query Syntax is constructed for you when you click on certain buttons. In a Command Based Query you use the syntax explicitly by typing each operator along with your search terms.

The Query Syntax uses a combination of standard operators used across the world and operators specific to Perceptive Enterprise Search. It is a useful set of tools for narrowing a search to a very specific range of words or phrases. The Query Syntax consists of:

Boolean operators
Positional operators
Precedence of operators and use of parentheses
Range searching
Date searching
Special symbols
Punctuation and special words
Filtering operators

Boolean Operators

These operators are used to find the existence of terms within a document in relation to other terms.

AND	The documents retrieved must contain both the words/phrases you have typed. The terms can appear anywhere in the document unless modified by additional operators.
OR	The document must contain at least one of the entered words or phrases. Both terms may be present.
NOT	The retrieved documents must contain the first term and must not contain the second, no matter where else it appears in the document.
XOR	The documents must contain either the first search term or the second but not both terms.
EXCEPT	The retrieved documents must contain the first search term but only if the second term is not in the same paragraph as the first. Both terms can appear in the document but they cannot appear within the same paragraph.
BUTNOT	The BUTNOT operator is an even more precise form of EXCEPT. The document must contain the first term and must not contain the second term immediately following the first. For example, the query "WORLD BUTNOT WORLD BANK" will find documents that contain the word "WORLD" and which may or may not contain "BANK" but will not find "WORLD BANK" where it appears as a phrase.

Positional Operators

These operators specify the location of each of your search terms in relation to each other within a document.

PHRASE		To find words which occur next to each other as a phrase just type the words as you would normally. In Perceptive Enterprise Search syntax there is no need to enclose phrases in quotes.
FAR FOLLOW	...	The retrieved documents must contain both terms and the second term must follow the appearance of the first. The second term can be anywhere in the document as long as it occurs after the first.
CLOSE FOLLOW	..	The retrieved documents must contain both search terms occurring in pairs. That is, the second term must immediately follow the first if a document is to be retrieved, without a repeated occurrence of the first term. If the first term occurs multiple times before the second term then it is the inner-most pair which is selected.
IN		The retrieved documents must contain the first term in a paragraph beginning with the second.
LABEL		Retrieved documents must contain a paragraph beginning with the specified search term.
IN PARAS	\\	Both search terms must occur within the specified number of paragraphs of each other. Placing a number between the lines indicates the number of paragraphs separating each term. No number between the \\ means the paragraphs must be consecutive. Single number (e.g. \5\). Search terms must occur within 5 paragraphs of each other. Number separated by comma (e.g. \5,10\). The second search term must appear within 5 to 10 paragraphs of the first. Negative numbers (e.g. \-5, +10\) indicate the paragraphs may begin before the appearance of the term. That is, the second term may appear up to 5 paragraphs before and up to 10 after the first.
NEAR BY	//	Both search terms must appear within the same paragraph. The order of appearance within the paragraph is not relevant. Placing a number between the lines indicates the number of words separating phrases. Single number (e.g. /5/) (this can also be written as W/5). The two search terms must appear within five words of each other. Order of appearance is not relevant. Number separated by comma (e.g. /5,10/). The search terms must appear in no fewer than 5 words and no more than 10 words of each other. Negative numbers (e.g. /-5, +10/). The search terms may appear from 5 words before to 10 words after each other.

Precedence of operators and use of parentheses

With queries which use more than one search term certain words and operators take precedence over others. Some operators will always be read by Perceptive Enterprise Search before others, regardless of where they occur in your query. The order of precedence is as follows:

Phrases
//, LABEL, IN, EXCEPT, TO, BEFORE, AFTER
\\
... , .. , AND, OR NOT

Where a query expression, or part thereof, contains several operators of the same precedence, Perceptive Enterprise Search will process them from left to right.

Use of parentheses

Parentheses (or brackets) allow you to control and define the order of precedence in complex queries as their use overrides the Perceptive Enterprise Search default searching order. They are not necessary in searches of two terms. However, when using queries with three or more terms parentheses can be invaluable in ensuring Perceptive Enterprise Search does exactly what you want it to.

Examples:

car allowance // manager OR salesperson

Using Query's default sequence, this will be interpreted as follows:

((car allowance) // manager) OR salesperson

...which means that documents need only to contain the word "salesperson" to be included with the more detailed search of "car allowance in the same paragraph as manager".

To change the query to "find the phrase car allowance in the same paragraph as manager or salesperson", use the following search:

car allowance // (manager OR salesperson)

Range Searching

When searching for numerical values within an index these operators allow you to specify a range values within which to search.

Specifies an alphabetic, numeric or date range to be searched. For example:

A123 TO A456

...will find "A123, A124, A125" ... and so on to "A456".

Note: for the TO operator to be used for dates the Intelligent Date Handling options must be selected for the index.

Search for a numeric range greater than or equal to the specified value. For example:

GE 1000

...will find "1000, 1001, 1002, 1003 ... 59001"

Search for a numeric range less than or equal to the specified value. For example:

LE 1000

...will find "1000, 999, 998, 997 ... -59001"

For increased accuracy when searching for numeric values select the Intelligent Number Handling option when creating your index.

Date Searching

Dates may be used as a search term if Intelligent Date Handling has been selected.

With Intelligent Date Handling set dates within your index will be found regardless of the way in which they are expressed in the documents. For example "February 10 2001", "2/10/01" and "Saturday, February 10, 2001" are all read as the same thing by Perceptive Enterprise Search. Relative or implied dates such as "March this year" or "two weeks from next Thursday" are not supported.

Date searching operators are:

AFTER	Search for documents containing dates falling after a specific date.
BEFORE	Search for documents containing dates falling before a specific date.
TO	Search for a range between and including the specified dates.

Special Symbols

Certain symbols can be used in place of letters or characters in query words to help refine or expand a query. Perceptive Enterprise Search will replace the symbols with letters according to its built-in logic and run the query on any resultant words. Using these symbols can save you having to think up and type all the permutations of a word that you need to include in a query.

Note that using these word expansion systems can still only retrieve words that are listed in your index.

*	This is a wildcard symbol and can be used to search for any number of significant characters. End of word Will find all words that start with the letters appearing before the symbol. For example 'manage' will find 'management', 'manages', 'managed', 'manager', etc. Middle of word* Will find all words that begin and end with the specified letters. e.g. 'ding' will find 'dying', 'divorcing', 'dividing', 'droning', etc. Beginning of word* Will find all words that end with the specified letters. Note that this option may significantly increase the search time required. Alone This will select the first word in all documents; a way of providing a list of all the documents in the index.
?	This is a single-character wildcard symbol and can be used to search for any single character. For example, searching for "h?t" would match "hat" or "hit" but not "heat". Both forms of wildcard may be used together in any combination desired, e.g."un*gra?ing".
~	Tense conflation (tense expansion). Asks Perceptive Enterprise Search to find all tense forms of the word specified. Trailing conflation Expands the tenses at the end of a word. For example, 'posting~' will find 'posted', 'posts', 'post' and 'poster', but not 'posture', 'postpone', etc. Leading conflation Expands the tenses at the beginning of a word. For example, '~post' would also find 'outpost', 'prepost' and 'repost'. By default Perceptive Enterprise Search automatically applies limited tense conflation to search terms. You can control tense expansion settings for your site by customizing the search page. If automatic tense conflation is on in the options, using this character turns it off; if automatic conflation is off, this character turns it on.
+	Toggle synonym expansion and thesaurus on or off. This symbol reverses the automatic synonym expansion setting in the search page. By default synonym expansion is set to on. Using this symbol in a query will turn it off for the word to which it is applied. For more information about Synonyms see Synonym Rings. If neither synonym expansion or the thesaurus are turned on in Options the + symbol will turn them both on for a given search term. If both are turned off in Options the + symbol will turn them both on. If only one option is active in Options the + symbol will switch it off (and leave the other option as it is) for a given search term.

Punctuation and special words

Punctuation

Perceptive Enterprise Search treats punctuation in a query in one of three ways:

Significant
This means the punctuated search word or phrase will match only with words or phrases containing matching punctuation characters.
Insignificant
Perceptive Enterprise Search ignores the occurrence of specified punctuation in the text, meaning that a search will bring up matching words and phrases whether punctuated or not.
Separators
Some punctuation marks will not be recognized as such by Perceptive Enterprise Search and treated the same way as spaces between words. For example, 'over-charge' will be indexed as two words - 'over' and 'charge'. By default the period, or full stop, is a punctuation character and thus not indexed. However, Perceptive Enterprise Search can be configured to index a period when it is embedded within a word.

You can change the configuration of punctuation in an index using Character Options.

Reserved Words

Reserved words are those used as Perceptive Enterprise Search operators. For example: AND, OR, NOT, LABEL and EXCEPT. If you need to include a reserved word in a query, precede the word with an insignificant character such as an underscore (_). This is a signal to Perceptive Enterprise Search to include the word as ordinary text in the search and not as an operator. For more information about significant and insignificant characters see the Character Options.

The // and \\ operators can be included in a search, if necessary, by enclosing them in quotes (i.e."//" and "\\").

Common Words

Common words are words such as I, THE, IT, IS and other words that frequently appear in many documents but do not frequently have a significant meaning. They are excluded from searches in order to save indexing space.

If your search includes a common word Perceptive Enterprise Search will remove it from the search and reconstruct the search around the gap. For example, 'beside the road' becomes 'beside /2,2/ road' automatically.
Where the common word includes a Boolean or positional operator the word is ignored. For example, 'manager and the' becomes simply 'manager'.
You will need to amend the Common Word list within Utilities if you wish to perform searches using words Perceptive Enterprise Search automatically defines as common. See Utilities for more information.

Filtering operators

Perceptive Enterprise Search includes the ability to filter results based on file name, file path, file date or format. The easiest way to do this is by clicking on the filter option and filling in the form. However, you can also include filters in your query syntax. The filter keywords are global to the entire query so it doesn't matter where you place them. They must be followed by a quoted parameter. For example, FNAMELIKE "*.TXT". If you wish to specify a list of values delimit each with a semi-colon. For example: FNAMELIKE "*.TXT;*.DOC".

FNAMELIKE	The file name must match the wild card specification in the quoted parameter which follows.
FNAMEUNLIKE	The file name must not match the wild card specification in the quoted parameter which follows.
PATHCONTAINS	The full path name must contain the string in the quoted parameter which follows.
PATHOMITS	The full path name must not contain the string in the quoted parameter which follows.
CATEGORYLIKE	The document category must match the wild card specification in the quoted parameter which follows.
CATEGORYUNLIKE	The document category must not match the wild card specification in the quoted parameter which follows.
FILEDATEBEFORE	The file date as stored on disk must precede the date specified in the quoted parameter which follows. The date format must be YYYYMMDD. For example: CAT FILEDATEBEFORE "20031225".
FILEDATEAFTER	Similar to FILEDATEBEFORE but the file date as stored on disk must follow the one given.
INDEXEDBEFORE	The document must have been indexed prior to the date specified in the quoted parameter which follows. The date format must be YYYYMMDD. For example: CAT INDEXEDBEFORE "20031225".
INDEXEDAFTER	Similar to INDEXEDBEFORE but the document must have been indexed after the date given.
FIRSTDATEAFTER	The first date appearing in the text of the document must follow the one given.
FIRSTDATEBEFORE	The first date appearing in the text of the document must precede the one given.