Word Hacks [Electronic resources]

Hack 82 Perform Power Text Searches with Regular Expressions

When wildcards just aren't enough, tap VBScript for powerful string searching in Word.

Although Word's wildcard searching is much better than most users realize, if you've previously used a language like Perl, Python, or JavaScript, you might prefer sticking with the special characters you already know for your searches. Besides, sometimes wildcards just aren't up to the job.

To borrow an example from O'Reilly's Learning Python, suppose you need to replace any occurrence of "red pepper" or "green pepper" with "bell pepper" if and only if they occur together in a paragraph before the word "salad," but not if they are followed (with no space) by the string "corn." That's definitely way out of Word's wildcards' league. (The pattern is \b(red|green)(\s+pepper(?!corn)(?=.*salad)), for those of you too impatient to wait until the full example at the end of this hack.)

Though VBA doesn't have built-in support for regular expressions, Microsoft does include a RegExp object with VBScript. With a slight change to your settings in the Visual Basic Editor, you can use the RegExp object in your macros.

First, select ToolsMacroVisual Basic Editor, and then choose ToolsReferences. In the next dialog, shown in Figure 9-3, check the "Microsoft VBScript Regular Expressions 5.5" box and click the OK button.

Now you can include instances of the RegExp object in your macros. The following section describes the RegExp object.

9.4.1 RegExp's Propeties and Methods

The RegExp object has four properties, described in the following list:

Pattern

The pattern string to search for.

Global

Whether search is for all occurrences that match Pattern, or just the first. This is a Boolean value, and the default is False.

IgnoreCase

Whether search is case-sensitive. This is a Boolean value, and the default is False.

MultiLine

Whether Pattern is matched across line breaks. This is a Boolean value, and the default is False.

The RegExp object has three methods, described in the following list:

Execute

Returns a Matches collection containing the matched substrings and information about those substrings.

Replace

Replaces all the substrings in a searched string that match a pattern with a replacement string. The syntax for this method is:

RegExpobject.Replace("string to search", "replacement pattern")

Test

Whether a search has successfully matched a pattern. Returns a Boolean value. Since this method always returns True if there were one or more successful matches, there's no need to set the Global property when using this method.

The Matches collection returned by the Execute method contains one or more Match objects, which have three properties, shown in the following list:

FirstIndex

The position of the Match's first character within the search string

Length

The number of characters in the Match

Value

The matched string

9.4.2 Using the RegExp Object in a Macro

The following macro interactively tests search patterns against the selected text.

Place this macro in the template of your choice [Hack #50] and either run it from the ToolsMacroMacros dialog or put a button for it on a menu or toolbar [Hack #1].

Sub RegExpTest( )
Dim re As RegExp
Dim strToSearch As String
Dim strPattern As String
Dim strResults As String
Dim oMatches As MatchCollection
Dim oMatch As Match
strToSearch = Selection.Text
Set re = New RegExp
re.Global = True
re.IgnoreCase = True
Do While (1)
strPattern = InputBox("Enter search pattern string:", _
"RegExp Search", ")
If Len(strPattern) = 0 Then Exit Do
re.Pattern = strPattern
Set oMatches = re.Execute(strToSearch)
If oMatches.Count <> 0 Then
strResults = Chr(34) & strPattern & Chr(34) & _
" matched " & oMatches.Count & " times:" _
& vbCr & vbCr
For Each oMatch In oMatches
strResults = strResults & _
oMatch.Value & _
": at position " & _
oMatch.FirstIndex & vbCr
Next oMatch
Else
strResults = Chr(34) & strPattern & Chr(34) & _
" didn't match anything. Try again."
End If
MsgBox strResults
Loop
End Sub

When you run this macro, you'll be prompted with the dialog shown in Figure 9-4.

The dialog shown in Figure 9-5 displays the search results.

The RegExp object supports the same metacharacters you might have seen in Perl:

\ | ( ) [ { ^ $ * + ? .

You also get all the classic Perl character-class shortcuts:

\d \D \s \S \w \W

For a full listing of special characters for using the RegExp object, see http://msdn.microsoft.com/library/default.asp?url=/library/en-us/script56l/vspropattern.asp.

9.4.3 Performing Replacements

When using the Replace method, you can group and reuse parts of the matched pattern. Known as backreferencing, this is a powerful technique. The following code snippet demonstrates how to change the format of some dates in a string:

re.Replace("(September) (\d\d?), (\d{4})", "$2 $1, $3")

This code will change a date like "September 12, 1978" into "12 September, 1978." Modifying the code to replace September with a different month won't require making any change to the replacement string, thanks to backreferencing.

Bringing all of this together, the following macro shows you how to use the "bell pepper" pattern discussed at the beginning of this hack to get the results shown in Figure 9-6.

Place this macro in the template of your choice [Hack #50] and either run it from the ToolsMacroMacros dialog or put a button for it on a menu or toolbar [Hack #1] :

Sub FixPeppers( )
Dim re As RegExp
Dim para As Paragraph
Dim rng As Range
Set re = New RegExp
re.Pattern = "\b(red|green)(\s+pepper(?!corn)(?=.*salad))"
re.IgnoreCase = True
re.Global = True
For Each para In ActiveDocument.Paragraphs
Set rng = para.Range
rng.MoveEnd unit:=wdCharacter, Count:=-1
rng.Text = re.Replace(rng.Text, "bell$2")
Next para
End Sub