Sounds Like Tokenizer 1.1 - C#, VB.NET and VB6 Code Samples

General Usage

The SoundsLike Phonetic Tokenizer for .Net applications is a phonetic normalization function that converts your customer names (or any text) into a numeric value that represents how it's pronounced.

By querying against these numbers, you will enrich your database searches and your applications will get around the problem of finding data that has different spellings or finding data that has similar or identical pronunciation.

Copy the SoundsLike.dll and supporting SoundsLike.xml document (for intellesense) to your projects' BIN directory and add a Reference.

TokenMaker objToken = new TokenMaker("*--- license key here ---*");

objToken.ProcessingFirstName = True;
objToken.StringToTokenize = "James";
if (objToken.IsValid) {
  myNarrow = objToken.TokenNarrow;
  myWide = objToken.TokenNarrow;
  myGroup = objToken.DiminutiveGroup;
}
else {
  strError = objToken.ErrorMessage;
  //Handle bad strings here
}

VB.NET

Dim objToken As New SoundsLike.TokenMaker("*--- license key here ---*")
Dim myNarrow As Integer
Dim myWide As Integer
Dim myGroup As Integer

objToken.ProcessingFirstName = True
objToken.StringToTokenize = "James"

If objToken.IsValid Then
  myNarrow = objToken.TokenNarrow
  myWide = objToken.TokenNarrow
  myGroup = objToken.DiminutiveGroup
Else
  strError = objToken.ErrorMessage
  'Handle bad strings here
End If

TokenMaker "EqualsNarrow" Method

This function may be used if the TokenNarrow property is not stored in the database. It can also be used for real-time compares of two words. If you set ProcessingFirstName to 'True', diminutive checking will be enabled and you will get 'True' result for 'Dick' = 'Richard' even though they do not sound alike.

//Initialize
TokenMaker objToken = new TokenMaker("*--- license key here ---*");
objToken.ProcessingFirstName = False;
   :
   :
//Loop processing
if (objToken.EqualsNarrow("yacht", "yawt")) {
  displayRecord = true;
}
else {
  displayRecord = false;
}
   :
   :

VB.NET

'Initialize
Dim objToken As New SoundsLike.TokenMaker("*--- license key here ---*")
objToken.ProcessingFirstName = False
   :
   :
'Loop processing
If objToken.EqualsNarrow("diaphragm", "diafram") Then
  displayRecord = True
Else
  displayRecord = False
End If
   :
   :

TokenMaker "EqualsNickname" Method (new for version 1.2)

This function may used to check if two names match because they are in the same nickname group. Only the first word of each string will be used. The ProcessingFirstName property is ignored and assumed to be 'True'. No phonetic matching will occur.

TokenMaker objToken = new TokenMaker("*--- license key here ---*");

if (objToken.EqualsNickname("Robert Smith", "Bob Smith")) {
//Names match; do something here
}
else {
//Names do not match
}

VB.NET

Dim objToken As New SoundsLike.TokenMaker("*--- license key here ---*")

If objToken.EqualsNickname("Dick Barnes", "Richard Barnes") Then
'Names match; do something here
Else
'Names do not match
End If

Supplemental "SOUNDEX" Method (new for version 1.2)

SOUNDEX is a phonetic algorithm for indexing names by their sound when pronounced in English. The basic aim is for names with the same pronunciation to be encoded to the same string so that matching can occur despite minor differences in spelling. SOUNDEX is the most widely known of all phonetic algorithms.

coming soon

VB.NET

Dim objToken As New SoundsLike.TokenMaker("*--- license key here ---*")

Dim strSoundex As String

'Regular 4-byte length SOUNDEX token
strSoundex = objToken.sup_SOUNDEX("Antidisestablishmentarianism") 'returns A533

'Custom length SOUNDEX token
strSoundex = objToken.sup_SOUNDEX("Antidisestablishmentarianism", 6) 'returns A53322

Supplemental "Levenshtein Distance" Method (new for version 1.2)

Levenshtein Distance is a measure of the similarity between two strings. The distance is the number of deletions, insertions, or substitutions required to transform the first string into the second. The greater the Levenshtein Distance, the more different the strings are. This is also sometimes called "Edit Distance."

coming soon

VB.NET

Dim objToken As New SoundsLike.TokenMaker("*--- license key here ---*")

Dim distance As Integer

distance = objToken.sup_LevenshteinDistance("Boland", "Coland") 'returns 1

Supplemental "Probability Of Match" Method (new for version 1.2)

This method returns a "Probability Of Match" percentage. It compares portions of two strings and computes a percentage (does not use the Levenshtein Distance). This is great for comparing mailing addresses to determine householding. By packing the name, address, suite, city, state, zip and country into a single string separated by spaces, the following example below returned a 98.4% probability of match:

string1 = "Robert MacIntyre Q 5337 Likini St P/H 2806 Honolulu HI 96818 USA"

string2 = "Robert McIntyre P/H 2806 5337 Likini St Honolulu HI 96818 USA"

and we were able to remove duplicate data from our database. Very powerful.

coming soon

VB.NET

Dim objToken As New SoundsLike.TokenMaker("*--- license key here ---*")

Dim percent As Decimal

percent = objToken.sup_ProbabilityOfMatch("Boland", "Coland") 'returns 85.7142857142857

Calling the SoundsLike DLL from Visual Basic 6

(only available for version 1.2) Several steps are necessary to call .Net DLLs from VB6. These instructions assume you have the .Net Framework (at least 1.1) installed and Visual Studio .Net (so you can open a .Net command prompt.) Please follow these steps and be sure to contact us via the support page if you encounter any difficulties.

1	Extract the zip file and place the SoundsLike.DLL in the desired location.
2	Open a .Net Command Prompt window. Usually found under Windows at: "Start" > "All Programs" > "Microsoft Visual Studio .NET 2003" > "Visual Studio .NET Tools" > "Visual Studio .NET 2003 Command Prompt"
3	Change directory to where the SoundsLike.DLL is located.
4	Run this command from the Command Prompt: tlbexp SoundsLike.dll /out:SoundsLike.tlb
5	Run this command from the Command Prompt: regasm /tlb=SoundsLike.tlb SoundsLike.dll
6	Run this command from the Command Prompt: gacutil /if SoundsLike.dll
7	Open your VB6 project.
8	Open the References window by selecting "Project" > "References..." from the tool bar.
9	Select "SoundsLike Phonetic Tokenizer" and click "OK".

Visual Basic 6.0

Dim objToken As SoundsLike.TokenMaker
Set objToken = New SoundsLike.TokenMaker

objToken.LicenseKey = "*--- license key here ---*"

objToken.PreScrubbingEnabled = True
objToken.ProcessingFirstName = True

objToken.StringToTokenize = "Dick"

Dim slNarrowToken As Long
Dim slWideToken As Long
Dim slNicknameGroup As Long
Dim slErrorCode As Byte
Dim slErrorMessage As String
Dim slPOM As Double
Dim slSOUNDEX As String
Dim slDistance As Long

slNarrowToken = objToken.TokenNarrow
slWideToken = objToken.TokenWide
slNicknameGroup = objToken.DiminutiveGroup
slErrorCode = objToken.errorCode
slErrorMessage = objToken.ErrorMessage

slPOM = objToken.sup_ProbabilityOfMatch("Bollard", "Collard")

slSOUNDEX = objToken.sup_SOUNDEX("Antidisestablishmentarianism") 'returns A533
slSOUNDEX = objToken.sup_SOUNDEX("Antidisestablishmentarianism", 6) 'returns A53322

slDistance = objToken.sup_LevenshteinDistance("Bollard", "Collard")

If objToken.EqualsNarrow("diaphragm", "diafram") = True Then
  'The words sound like each other!
End If

If objToken.EqualsWide("yacht", "yawt") = True Then
  'The words sound like each other!
End If

If objToken.EqualsNickname("Robert", "Bob") = True Then
  'The first names match!
End If

Visual Basic 6 / SqlServer Case Study

(only available for version 1.2) Here is an example of how you can use the SoundsLike Tokenizer in a VB6 application. These instructions assume you have the .Net Framework (at least 1.1) installed. Please follow these steps and be sure to contact us via the support page if you encounter any difficulties.

Table Creation Script

CREATE TABLE [dbo].[tblTestPerson] (
[CustomerID] [int] IDENTITY (1, 1) NOT NULL ,
[Prefix] [varchar] (10) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[FirstName] [varchar] (20) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[MiddleName] [varchar] (15) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[LastName] [varchar] (35) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[Suffix] [varchar] (10) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[TokenFirstName] [bigint] NULL ,
[TokenLastName] [bigint] NULL ,
[TokenNickNameGrp] [bigint] NULL ,
[Credentials] [varchar] (50) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[Gender] [char] (1) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[DateOfBirth] [datetime] NULL ,
[SSN] [char] (11) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[PhotoURL] [varchar] (50) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[PhotoOnFile] [bit] NULL ,
[PhotoShownOutside] [bit] NULL ,
[SpouseName] [varchar] (50) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
[LastUpdated] [datetime] NULL ,
[LastUpdatedBy] [varchar] (15) COLLATE SQL_Latin1_General_CP1_CI_AS NULL
) ON [PRIMARY]
GO

ALTER TABLE [dbo].[tblTestPerson] WITH NOCHECK ADD
CONSTRAINT [PK_tblTestPerson] PRIMARY KEY CLUSTERED
(
[CustomerID]
) ON [PRIMARY]
GO

ALTER TABLE [dbo].[tblTestPerson] ADD
CONSTRAINT [DF_tblTestPerson_TokenFirstName] DEFAULT (0) FOR [TokenFirstName],
CONSTRAINT [DF_tblTestPerson_TokenLastName] DEFAULT (0) FOR [TokenLastName],
CONSTRAINT [DF_tblTestPerson_TokenNickNameGrp] DEFAULT (0) FOR [TokenNickNameGrp]
GO

Initial database load (pre-tokenizing existing data)

Private Sub btnIntialLoad_Click() 

  '---------------------------------------------------------------------------
  'Create an instance of the SoundsLike Phonetic Tokenizer
  '---------------------------------------------------------------------------
  Dim objToken As SoundsLike.TokenMaker
  Set objToken = New SoundsLike.TokenMaker 

  objToken.LicenseKey = "*--- license key here ---*"
  objToken.PreScrubbingEnabled = True 

  Dim slTokenFirstName As Long
  Dim slTokenLastName As Long
  Dim slNicknameGroup As Long
  
  '---------------------------------------------------------------------------
  'Define and establish database connection
  '---------------------------------------------------------------------------
  Dim objConn As ADODB.Connection
  Dim objConnUpd As ADODB.Connection
  Dim objRs As ADODB.Recordset
  Dim strSQL As String 

  Set objConn = CreateObject("ADODB.Connection")
  Set objConnUpd = 
   CreateObject("ADODB.Connection")  Set
  
  objRs = CreateObject("ADODB.Recordset")
  objConn.ConnectionString= _
      "Driver={SQLServer};Server=SNDLK;Database=dbCustomer;UID=xxx;PWD=xxx;"
  objConn.CommandTimeout = 90 ' seconds
  objConn.ConnectionTimeout = 45 ' seconds
  
  objConn.Open
  objRs.ActiveConnection = objConn 

  objConnUpd.ConnectionString = objConn.ConnectionString
  objConnUpd.Open
  
  '---------------------------------------------------------------------------
  'Build SQL
  '---------------------------------------------------------------------------
  objSQL = "SELECT CustomerID, FirstName, LastName " _
         & "FROM dbo.tblTestPerson " _
         & "WHERE (TokenFirstName < 1) OR (TokenLastName < 1)"
          
  '---------------------------------------------------------------------------
  'Execute SQL
  '---------------------------------------------------------------------------
  Set objRs = CreateObject("ADODB.Recordset")
  objRs.ActiveConnection = objConn
  objRs.Open objSQL
    
  '---------------------------------------------------------------------------
  'Loop thru and tokenize all existing names already in the database
  ' ===This is a one-time operation to pre-tokenize existing data===
  '---------------------------------------------------------------------------
  While Not objRs.EOF
  
    'Tokenize the first name and get the nickname group (if any)
    objToken.ProcessingFirstName = True
    objToken.StringToTokenize = objRs("FirstName")
    slTokenFirstName = objToken.TokenNarrow
    slNicknameGroup = objToken.DiminutiveGroup
    If slNicknameGroup = -1 Then slNicknameGroup = 0 'Use 0 to indicate no nickname
  
    'Tokenize the last name
    objToken.ProcessingFirstName = False
    objToken.StringToTokenize = objRs("LastName")
    slTokenLastName = objToken.TokenNarrow
    
    strSQL = "Update tblTestPerson " _
           & "Set TokenFirstName = " & slTokenFirstName & "," _
           & "    TokenNickNameGrp = " & slNicknameGroup & "," _
           & "    TokenLastName = " & slTokenLastName & " " _
           & "Where CustomerID = " & objRs("CustomerID")
  
    objConnUpd.Execute strSQL
        
    objRs.MoveNext
  Wend
  
  '---------------------------------------------------------------------------
  'Clean up
  '---------------------------------------------------------------------------
  objRs.Close
  objConn.Close
  objConnUpd.Close 

  Set objRs = Nothing
  Set objConn = Nothing
  Set objConnUpd = Nothing
  Set objToken = Nothing 

End Sub

Processing the search in VB6

Private Sub btnSearch_Click() 

  '---------------------------------------------------------------------------
  'Create an instance of the SoundsLike Phonetic Tokenizer
  '---------------------------------------------------------------------------
  Dim objToken As SoundsLike.TokenMaker
  Set objToken = New SoundsLike.TokenMaker 

  objToken.LicenseKey = "*--- license key here ---*"
  objToken.PreScrubbingEnabled = True
  objToken.ProcessingFirstName = True 

  Dim slTokenFirstName As Long
  Dim slTokenLastName As Long
  Dim slNicknameGroup As Long
  
  '---------------------------------------------------------------------------
  'Tokenize the first name the user is looking for
  '  and get the nickname group (if any)
  '---------------------------------------------------------------------------
  objToken.StringToTokenize = tbFirstName.Text
  If Not objToken.IsValid Then MsgBox (objToken.ErrorMessage) 'Testing
  slTokenFirstName = objToken.TokenNarrow
  slNicknameGroup = objToken.DiminutiveGroup
  
  '---------------------------------------------------------------------------
  'Tokenize the last name the user is looking for
  '---------------------------------------------------------------------------
  objToken.ProcessingFirstName = False
  If Not objToken.IsValid Then MsgBox (objToken.ErrorMessage) 'Testing
  objToken.StringToTokenize = tbLastName.Text
  slTokenLastName = objToken.TokenNarrow


  '---------------------------------------------------------------------------
  'Define and establish database connection
  '---------------------------------------------------------------------------
  Dim objConn As ADODB.Connection
  Dim objRs As ADODB.Recordset
  Dim strSQL As String 

  Set objConn = CreateObject("ADODB.Connection")
  Set objRs = CreateObject("ADODB.Recordset")
  
  objConn.ConnectionString = _
       "Driver={SQL Server};Server=SNDLK;Database=dbCustomer;UID=xxx;PWD=xxx;"
  objConn.CommandTimeout = 90 ' seconds
  objConn.ConnectionTimeout = 45 ' seconds
  
  objConn.Open
  objRs.ActiveConnection = objConn 

  '---------------------------------------------------------------------------
  'Build SQL - THIS IS HOW IT'S DONE RIGHT HERE > > >
  '  look for text matches AND numeric token matches
  '---------------------------------------------------------------------------
  objSQL = "SELECT FirstName, MiddleName, LastName, " _
         & "PhotoURL, PhotoOnFile, PhotoShownOutside, SpouseName " _
         & "FROM dbo.tblTestPerson " _
         & "WHERE  ( (FirstName = '" & Trim(tbFirstName.Text) & "') " _
         & "OR (TokenFirstName = " & slTokenFirstName & ") " _
         & "OR (TokenNickNameGrp = " & slNicknameGroup & ") ) " _
         & "AND    ( (LastName = '" & Trim(tbLastName.Text) & "') " _
         & "OR (TokenLastName = " & slTokenLastName & ") ) " _
         & "ORDER BY LastName, FirstName"
          
  '---------------------------------------------------------------------------
  'Execute SQL
  '---------------------------------------------------------------------------
  Set objRs = CreateObject("ADODB.Recordset")
  objRs.ActiveConnection = objConn
  objRs.Open objSQL
    
  '---------------------------------------------------------------------------
  'Process results
  '---------------------------------------------------------------------------
  While Not objRs.EOF
    'Load search result screen here
    objRs.MoveNext
  Wend
  
  '---------------------------------------------------------------------------
  'Clean up
  '---------------------------------------------------------------------------
  objRs.Close
  objConn.Close
  Set objRs = Nothing
  Set objConn = Nothing
  Set objToken = Nothing 

End Sub

Don't forget to tokenize new Customer data when it gets added.

Frequently Asked Questions