GET TEXT KEYWORDS

Home

4D v16.3

GET TEXT KEYWORDS


GET TEXT KEYWORDS ( text ; arrKeywords {; *} )

Parameter	Type	Description
text	Text	Original text
arrKeywords	Text array	Array containing keywords
*	Operator	If passed = unique words

Description

The GET TEXT KEYWORDS command splits all the text into individual words and creates an item in the arrKeywords text array for each word.

4D uses the same algorithm to break up text into individual words that it does to build a Keywords index. It is based on the ICU library. For more information about how text is separated into words, refer to the following address: http://userguide.icu-project.org/boundaryanalysis.

Note: At the request of users, we introduced an exception for French and Italian: the apostrophe (') followed by either a vowel or the letter "h" is considered as a word separator. For example, the strings "L'homme" or "l'arbre" are split into "L’"+"homme" and "l'"+"arbre".

The algorithm used differs according to whether or not the Consider only non-alphanumeric chars for keywords option is checked in the Database settings (refer to Database/Data storage page in the Design Reference manual).

In the text parameter, pass the original text to be split into words. This can be styled text, in which case style tags are ignored.

For the arrKeywords parameter, the command fills this text array with the words extracted from the text.

If you pass the optional * parameter, the command only stores each different keyword once in arrKeywords. By default, if this parameter is omitted, all the words extracted from the text are stored in the array, even when they appear more than once.

This command gives you a simple way to search records containing large amounts of text with the assurance of using the same keywords as 4D. For example, imagine you have a text containing "10,000 Jean-Pierre BC45". If this text is split into the keywords "10,000" + "Jean" + "Pierre" + "BC45", then the array contains 4 elements. Then it is easy to loop through this array by programming to find records containing one or more of these keywords using the % operator (see examples).

Example 1

In a form with a search area, users can enter one or more word(s). When a user validates this form, we look for records where the MyField field contains at least one of the words entered by the user.

  // vSearch is the variable of the search area in the form

 GET TEXT KEYWORDS(vSearch;arrSearch;*)

  //* in case a user enters the same word more than once

 CREATE SET([MyTable];"Totalfound")

 $n:=Size of array(arrSearch)

 For($i;1;$n)

    QUERY([MyTable];[MyTable]MyField % arrSearch{$i})

    CREATE SET(([MyTable];"found")

    UNION("Totalfound";"found";"Totalfound")

 End for

 USE SET("Totalfound")

Example 2

In the same form as before, we look for records where the MyField field contains all the words entered by the user.

  // vSearch is the variable of the search area in the form

 GET TEXT KEYWORDS(vSearch;arrSearch;*)

 $n:=Size of array(arrSearch)

 QUERY([MyTable];[MyTable]MyField >=0;*)

  // initializing search = all records

 For($i;1;$n)

    QUERY([MyTable];&;[MyTable]MyField % arrSearch{$i};*)

  // add criterion

 End for

 QUERY([MyTable]) //search

Example 3

To count words in a text:

 GET TEXT KEYWORDS(vText;arrWords) // all words

 $n:=Size of array(arrWords)

 GET TEXT KEYWORDS(vText;arrWords;*) // different words

 $m:=Size of array(arrWords)

 ALERT("This text contains "+String($n)+" separate words among "+String($m))