TriSys provides a powerful and flexible free-text search facility able to find any word or any combination of words.
This is achieved through the use of a ‘Search Index’.
An index in a book allows the reader to rapidly locate required information, providing a list of words/topics cross-referenced to page numbers.
In the same way TriSys periodically (each night, lunchtime etc.) checks through every word (except ‘common’ words – see below in A.2.7) in every document associated with database records and builds an index; i.e. a table containing details of where all the different words are within the system’s documents.
Thus when a free-text search expression is entered, TriSys can very quickly locate the required words using the index.
a wide variety of ‘Operators’ (e.g. AND, OR, NOT) that can be used to form ‘Search Expressions’, precisely reflecting the information that is to be found.
e.g. COBOL AND CICS AND DB2 AND DBA
The simplest search expression is a single search criterion. A more advanced search expressions consist of many terms separated by an operators, as above.
There are four kinds of operators in TriSys, details of their use now follows:
Note: The operators in the following text are emboldened and capitalised for the sake of clarity. It is not necessary to embolden or capitalise them when using the Free Text search.
|
Operator
|
Use |
|
AND |
Match both criteria, e.g.:
ORACLE AND UNIX
finds all documents containing both ORACLE and UNIX.
|
|
OR |
Match either criteria, e.g.:
LOTUS 1-2-3 OR EXCEL
finds documents containing LOTUS 1-2-3 or EXCEL. Documents containing both are also included.
|
|
NOT |
Match the first but not the second criteria, e.g.:
IBM NOT FUJITSU
finds documents containing IBM, but excludes any containing FUJITSU.
|
|
XOR |
Match either the first or second criteria, but not both. E.g.:
ORACLE XOR SYBASE
finds documents containing ORACLE, and those containing SYBASE, but excludes any which contain both.
|
|
EXCEPT |
Match the first criteria but only if the second is not in the same paragraph as the first, e.g.:
COBOL EXCEPT BULL
finds documents which contain COBOL, but excludes those containing BULL in the same paragraph. |
More than one Boolean operator can be used when specifying selection criteria. To avoid ambiguity, enclose each pair of criteria and their Boolean operator in parenthesis to include the logical manner in which you wish TriSys to interpret your criteria. E.g.:
(IBM AND COBOL) OR (DIBIL AND ICL)
Positional operators are used to specify the relative position of words and phrases, and the order in which they occur.
|
Operator
|
Specifies |
|
… (FAR FOLLOW) |
Both criteria must appear with the second following the first, e.g.:
COBOL … 4GL
finds the first occurrence of COBOL, the last occurrence of 4GL and all occurrences in between.
|
|
.. (CLOSE FOLLOW) |
The criteria must appear in pairs, e.g:
ORACLE .. INGRES
finds the first occurrence of ORACLE, the first occurrence of INGRES and all the other occurrences of the two words, providing they appear in pairs.
|
|
IN |
The first criterion must be contained in a paragraph that starts with the second creation, e.g.:
DEGREE IN UNIVERSITY
finds all occurrences of DEGREE in paragraphs which begin with UNIVERSITY. The second term, known as the label, must not begin with a common word.
|
|
LABEL |
The document must contain a paragraph that starts with the criterion, e.g.:
LABEL ENTHUSIASTIC SPORTSPERSON
finds paragraphs which begin with ENTHUSIASTIC SPORTSPERSON. A paragraph label must not begin with a common word.
|
|
\\ (IN PARAS) |
Both criterion must occur within the specified number of paragraphs, e.g.:
SUPERVISOR \\ MANAGER
Finds documents that contain SUPERVISOR and MANAGER within one paragraph of each other.
You can specify the positioning as follows:
Entering \4\, locates documents containing MANAGER within four paragraphs of SUPERVISOR.
Entering \2,5\, locates documents containing MANAGER within two to five paragraphs of SUPERVISOR.
Entering \-3, +3\, locates documents containing MANAGER up to three paragraphs before SUPERVISOR and up to three paragraphs after it.
|
|
// (NEAR BY) |
Both criterion must appear in the same paragraph, e.g.:
YEAR END // ACCOUNTS
finds documents where both YEAR END and ACCOUNTS appear in the same paragraph, regardless of order.
Numbers may be used as indicated above in ‘IN PARAS’ to define criteria which are a certain distance apart. |
Use the following operators to specify; alphabetic, numeric or date ranges.
|
Operator |
Specifies |
|
TO |
The document must contain a word within a specified range, e.g.:
A20 TO A25
finds the following: A20, A21, A22, A23, A25, A26.
|
Use the following operators to search for dates:
|
Operator |
Specifies |
|
AFTER |
The document must contain a date after the date specified, e.g.:
AFTER 1 Jan 1997
|
|
BEFORE |
The document must contain a date before the date specified, e.g.:
BEFORE 31 Dec 1998 |
A.2.5 Precedence of operators – order of processing.
Search criteria are processed in a certain order governed by rules known as ‘precedence’.
The order in which types of operator are processed is indicated below.
Phrases
//, LABEL, IN, EXCEPT, TO, BEFORE, AFTER
\\
…, .., AND, OR, NOT
Where a search expression contains several operators of the same precedence, they are processed from left to right. E.g.:
UNIX AND C OR C++
finds documents which contain either both UNIX and C, or just C++.
Note: C++ must be entered in quotation marks because + is a search operator.
To find documents which contain C++ and UNIX or C, enter the queries:
Unix or c and 'c++'
A.2.6 How to use parenthesis in searches
Parentheses control the order in which search terms are processed, and hence the meaning, overriding the system precedence rules, e.g.:
When searching for a Hardware tester or QA specialist;
Locate documents with the word ‘hardware’ adjacent to ‘test*’ or ‘QA’ or ‘quality’.
Entering: Hardware //test or qa
Will locate documents containing ‘hardware’ adjacent to ‘test’, and those which contain just ‘QA’.
Whereas entering: HARDWARE // (TEST OR QA)
Will locate documents containing ‘hardware’ adjacent to ‘test’, and also those containing ‘hardware’ adjacent to ‘QA’.
You can use parenthesis elsewhere in searches as the following examples show:
(MANAGER OR SUPERVISOR) // BONUS
(SUPERVISOR OR TEAM LEADER) responsible for (ANNUAL OR FINAL) accounts
YEAR END … (BONUS OR AWARD)
Some words, such as ‘the’, ‘it’ and ‘is’ appear frequently in documents. To save index space, these words are defined as ‘common’ and are not indexed. If your search expression contains a common word, they are removed it from the search, which is then intelligently reconstructed around the gap to retain the sense of the original expression, e.g.:
MANAGE THE DEPARTMENT
TriSys searches on MANAGE/2,2/DEAPRTMENT replacing the common word the NEAR BY operator to ensure that DEPARTMENT is located within two words of MANAGED.
If your search criterion contains a word that is one of the operators described, for example, "and", TriSys Professional 4 ignores it.
TriSys keeps a list of significant and insignificant characters (those beyond the normal range of text characters, that is A to Z and the international character set).
Significant characters are included when a word is indexed, whereas insignificant characters are treated as if they do not exist in the text.
By default, the ‘-‘ is considered a punctuation character or word separator. E.g., over-qualified would be considered to be two words: the word ‘over’ followed by the word ‘qualified’. If the default is changed and the hyphen is defined s insignificant, then both the words containing ‘overqualified’ and ‘over-qualified’ would be considered as ‘overqualified’.
Soft hyphens used in many word processors are reorganised and a word containing such a character is always considered as a single word.
By default, the full stop is treated as a punctuation character and not normally indexed. However, TriSys can be configured so that full-stops are indexed when they are embedded within a word. E.g. you may wish to search for the term ABC.123.
TriSys treats all ‘operators’ covered in A.2 above as reserved words. To search for one of these, (e.g. to find occurrences of ‘LABEL’), the word must be prefixed with an insignificant character, (e.g. underscore ‘_’).
Are as follows:
|
* or ! |
Known as ‘Wildcards’, use either of these to search for specific text strings, which may be followed by any number of other significant characters, e.g. entering
MANAGE* (or MANAGE!)
finds all words that start with ‘MANAGE’; such as ‘manage’, ‘management’, ‘managers’, ‘managed’ etc.
Wildcard characters can be used at the beginning of a word; but as the wildcard character is placed further to the left in the string, the longer the search will take.
Entering ‘*’ on its own, produces a list of all the documents in the database, (as the first word in every document is selected).
Only one wildcard character per text string is allowed. |
|
|
|
? A wildcard representing any single significant character,
e.g.: entering bo?t
will find; ‘boat’, ‘bout’ etc.