Visual Basic Q&A

Symptoms
This sample demonstrates how a Visual Basic 4.0 application calls Fortranand Visual C DLL’s. Specifically, it shows how to pass fixed lengthstrings to and from those DLL’s.
Resolution
The following file is available for download from the Microsoft Download Center:
Vbstring.exe(http://download.microsoft.com/download/vb40ent/sample40/1/w9xnt4/en-us/vbstring.exe)Release Date: Jan-01-1997
For additional information about how to download Microsoft Support files, click the following article number to view the article in the Microsoft Knowledge Base:
119591?(http://support.microsoft.com/kb/119591/EN-US/)How to Obtain Microsoft Support Files from Online ServicesMicrosoft scanned this file for viruses. Microsoft used the most current virus-detection software that was available on the date that the file was posted. The file is stored on security-enhanced servers that help to prevent any unauthorized changes to the file.
Visual Basic 4.0 replaced the string management system used by previousversions of Visual Basic with a more robust string management system.Visual Basic 4.0 relies on the OLE automation data types, such as BSTR andsafe arrays, to manage string usage. However, passing arrays of numerictypes, integers, or reals were not changed. For instance, before passing anarray of strings to a DLL, the strings must be converted to an array ofbytes, and then converted back to strings on return from the DLL. Please,refer to VB4DLL.TXT that ships with Visual Basic 4.0 for more information.
All new C DLL’s should use the syntax described in the VB4DLL.TXT filebecause the API calls that are employed allow for improved error handling.The programmer can do something in the event of an error, such as display amessage box and abort the operation.
Please note that the new Fortran DLL’s cannot use the syntax described inVB4DLL.TXT because they have no direct access to the correct APIfunctions.
The sample also demonstrates different methods you can use to resolve thefunction symbol within the DLL. A mismatch in the function symbol resultsin a Visual Basic runtime error, “Specified DLL function not found (Error453).” The sample includes two methods to resolve C function symbols andthree methods to resolve Fortran function symbols.
Each source file contains more information pertinent to that language.
This sample also enables you to choose whether the data is passed to eitherthe C or Fortran DLL.
This sample shows how to pass the following:
Two dimensional array of 4-byte integers.Two dimensional array of 8-byte floating point numbers.String.One dimensional array of strings.Two dimensional array of strings.One dimensional array of 4-byte integers, a one dimensional array of8-byte floating point numbers, and a one dimensional array of strings.
Sample Files

FileNameDescription——————————————————-cdll.cC source file to build 32-bit DLLfordll.forFortran source file for FPS NT 1.0 referencefordll32.f90Fortran source file for the 32-bit DLLvbstring.vbpVisual Basic 4.0 project filevbstring.frmVisual Basic form including event handlers
Building DLLsThis sample assumes you are using the following development environments:
Microsoft Visual C++, version 4.0Microsoft FORTRAN PowerStation for Windows 95 and Windows NT, version 4.0You will need to setup the environment before running BUILDDLL.BAT.At an MS-DOS command-line prompt type:

C:\MSDEV\BIN\VCVARS32.BAT Build the C DLL. At the command-line prompt type:

BUILDDLL C Build the Fortran DLL. At the same prompt type:

BUILDDLL Fortran Start Visual Basic 4.0 and open the project file VBSTRING.VBP.Run the Visual Basic application by creating an executable file (.exe)or running the program from within Visual Basic.IMPORTANT: The DLLs must be in the \WINDOWS\SYSTEM directory, the directory containing the Visual Basic executable file, or one of the environment file paths.

Notes on Arrays in Different LanguagesGiven an upper bound of n, note that in Visual Basic all of the arrayranges are specified as “1 to n”. When a lower bound is not specified,Visual Basic, by default, assumes 0<=index<=n where “n” is inclusive. C array subscripts are assumed to be in the range 0<=index<=n-1. Fortran array ranges are 1<=index<=n. You need to be careful when you assign the index ranges, and make sure that they match. If you exceed them, you will probably get an Access Violation, an Unhandled Exception, or a random run-time error in your Visual Basic Application.

LanguageDeclarationRange# Items in Array—————————————-BasicDim X(10) As Integer0<=index<=1011Basic*Dim X(1 to 10) As Integer1<=index<=1010Cint X[10];0<=index<=910Fortraninteger X(10)1<=index<=1010Fortran*integer X(0:9)0<=index<=910* With lower bound specified. C is always zero-based.

Symptoms
Microsoft Access allows you to modify an existing field’s data type. To do so programmatically, Microsoft Jet 4.0 introduces the ALTER TABLE ALTER COLUMN DDL statement. However, there is no equivalent for Microsoft Jet 3.5.
This article demonstrates a method to alter a field’s data type using DAO objects.
Resolution
Modifying a field’s data type requires the following steps:Rename the old field.Add a new field.Copying the data from the old field to the new field.Delete the old field.If the table has any indexes or relations, the relationships and indexes must be dropped prior to performing the steps above, then re-established after completion of the steps above.
Microsoft Access handles indexes but not relationships when changing data types.
The Jet 4.0 ALTER TABLE ALTER COLUMN DDL statement has similar limitations.
The sample code provided handles both indexes and relationships.It also contains error handling to roll back the changes and report on any problems.
The main procedure is ChangeFieldType. It takes the following arguments:db – an open Database object where the table resides.TableName – the name of the table where the field resides.FieldName – the name of the field to be changed.NewType – the new data type for the field.NewAllowZeroLength – new value for the AllowZeroLength property.NewAllowNulls – used to set the Required property of the new field.NewAttributes – used to set the Attributes property of the new field.Note: This procedure is for illustration purposes only. For example, the procedure copies only basic field properties. In addition to these basic field properties, other field properties might also have to be copied. These additional field properties include ValidationRule, ValidationText, DecimalPlaces, and others, depending on the field type. In addition, the procedure does not copy user-defined properties.
The other procedures, RecordRelationInfo, RecordIndexInfo, IsField, and MakeArray, are helper procedures used by the main function.
Sample CodeThis sample changes the CustomerID field in the Customers table from a five character field to an eight character field.
The sample uses the Nwind database that comes with Visual Basic.
In Visual Basic, create a new Standard EXE project.
Form1 is created by default.Add a command button to Form1. Command1 is created by default.On the Project menu, select References.
In the References dialog, select the Microsoft DAO Object Library.On the Project menu, select Add Module to add a Code Module.
Module1 is created by default.Paste the following code into the General Declarations section of Module1’s Code Window:

Option Compare TextOption ExplicitConst CFT_Failed As Long = 55555Private Const R_NAME = 0, R_ATTRIBUTES = 1, R_TABLE = 2, R_FOREIGNTABLE = 3, R_FIELD = 4, R_FOREIGNFIELD = 5Private Const I_NAME = 0, I_PRIMARY = 1, I_UNIQUE = 2, I_REQUIRED = 3, I_IGNORENULLS = 4, I_CLUSTERED = 5, I_FIELD = 6, I_FIELDATTRIBUTES = 7Public Sub ChangeFieldType(db As Database, _ByVal TableName As String, _ByVal FieldName As String, _ByVal NewType As Integer, _Optional NewSize As Long, _Optional NewAllowZeroLength As Boolean = False, _Optional NewAllowNulls As Boolean = True, _Optional NewAttributes As Long)’ User-defined properties are not maintainedDim td As TableDef, I As Index, R As Relation, F As Field’ loop iterators for Indexes, Fields, and Relations collections:Dim I1 As Long, F1 As Long, R1 As LongDim colR As Collection, colI As CollectionDim E_Desc As String, Process As String, SubProcess As String, E As ErrorDim TempFieldName As String, Suffix As Long, OldName As StringDim Temp As VariantDim OrdinalPosition As LongSet colI = New CollectionSet colR = New CollectionOn Error GoTo CFT_ErrDBEngine(0).BeginTrans’ Enumerate relations and save/remove themDBEngine(0).BeginTransProcess = “Removing relations on [" & TableName & "]![" & FieldName & "]“SubProcess = “”For R1 = db.Relations.Count – 1 To 0 Step -1Set R = db.Relations(R1)If R.Table = TableName ThenFor F1 = 0 To R.Fields.Count – 1Set F = R.Fields(F1)If F.Name = FieldName ThenRecordRelationInfo R, colRSubProcess = “Removing relation ” & R.Namedb.Relations.Delete R.NameExit ForEnd IfNext F1ElseIf R.ForeignTable = TableName ThenFor F1 = 0 To R.Fields.Count – 1Set F = R.Fields(F1)If F.ForeignName = FieldName ThenRecordRelationInfo R, colRSubProcess = “Removing relation ” & R.Namedb.Relations.Delete R.NameExit ForEnd IfNext F1End IfNext R1Set F = NothingSet R = NothingDBEngine(0).CommitTrans’ Enumerate indices and save/remove themDBEngine(0).BeginTransProcess = “Removing indexes on [" & TableName & "]![" & FieldName & "]“SubProcess = “”db.TableDefs.RefreshSet td = db(TableName)td.Indexes.RefreshFor I1 = td.Indexes.Count – 1 To 0 Step -1Set I = td.Indexes(I1)If I.Foreign <> True ThenFor F1 = 0 To I.Fields.Count – 1Set F = I.Fields(F1)If F.Name = FieldName ThenRecordIndexInfo I, colISubProcess = “Removing index ” & I.Nametd.Indexes.Delete I.NameExit ForEnd IfNext F1End IfNext I1Set F = NothingSet I = NothingDBEngine(0).CommitTrans’ Rename FieldDBEngine(0).BeginTransProcess = “Renaming field”SubProcess = “”td.Fields.RefreshSet F = td(FieldName)OrdinalPosition = F.OrdinalPosition’ save this value’ determine a field name not in useSuffix = 0DoSuffix = Suffix + 1TempFieldName = “XXX” & SuffixLoop While IsField(td, TempFieldName)’ rename the fieldSubProcess = “to ” & TempFieldNameF.Name = TempFieldNameSet F = NothingDBEngine(0).CommitTrans’ Add new FieldDBEngine(0).BeginTransProcess = “Adding new field”SubProcess = “”td.Fields.RefreshSet F = td.CreateField(FieldName, NewType)If NewSize Then F.Size = NewSizeF.AllowZeroLength = NewAllowZeroLengthF.Required = Not NewAllowNullsF.Attributes = NewAttributesF.OrdinalPosition = OrdinalPositiontd.Fields.Append FSet F = NothingSet td = NothingDBEngine(0).CommitTrans’ Copy dataDBEngine(0).BeginTransProcess = “Copying data from ” & TempFieldName & ” to ” & FieldNameSubProcess = “”db.Execute “UPDATE [" & TableName & "] SET [" & FieldName & "]=[" & _TempFieldName & "]“, dbFailOnErrorDBEngine(0).CommitTrans’ Delete temporary fieldDBEngine(0).BeginTransProcess = “Deleting temporary field ” & TempFieldNameSubProcess = “”Set td = db(TableName)td.Fields.Delete TempFieldNameDBEngine(0).CommitTrans’ Add back IndicesDBEngine(0).BeginTransProcess = “Adding indexes back into table”SubProcess = “”Set td = db(TableName)td.Fields.Refreshtd.Indexes.RefreshOldName = “”Set I = NothingFor Each Temp In colIIf Temp(I_NAME) <> OldName ThenIf Not (I Is Nothing) Then’ handle first time through caseSubProcess = “Adding index ” & I.Nametd.Indexes.Append IEnd IfSet I = td.CreateIndex(Temp(I_NAME))I.Primary = Temp(I_PRIMARY)I.Unique = Temp(I_UNIQUE)I.Required = Temp(I_REQUIRED)I.IgnoreNulls = Temp(I_IGNORENULLS)I.Clustered = Temp(I_CLUSTERED)End IfSet F = I.CreateField(Temp(I_FIELD))F.Attributes = Temp(I_FIELDATTRIBUTES)’ to handle descending indexI.Fields.Append FNext TempIf Not (I Is Nothing) Then’ handle case of no indexesSubProcess = “Adding index ” & I.Nametd.Indexes.Append IEnd IfSet F = NothingSet I = NothingSet td = NothingDBEngine(0).CommitTrans’ Add back relationsDBEngine(0).BeginTransProcess = “Adding relations back into database”SubProcess = “”OldName = “”db.Relations.RefreshSet R = NothingFor Each Temp In colRIf Temp(I_NAME) <> OldName ThenIf Not (R Is Nothing) Then’ handle first time through caseSubProcess = “Adding relation ” & R.Namedb.Relations.Append REnd IfSet R = db.CreateRelation(Temp(R_NAME), Temp(R_TABLE), _Temp(R_FOREIGNTABLE), Temp(R_ATTRIBUTES))End IfSet F = R.CreateField(Temp(R_FIELD))F.ForeignName = Temp(R_FOREIGNFIELD)R.Fields.Append FNext TempIf Not (R Is Nothing) Then’ if there are no indexes…SubProcess = “Adding relation ” & R.Namedb.Relations.Append REnd IfSet F = NothingSet R = NothingDBEngine(0).CommitTrans’ Commit all pending chhangesDBEngine(0).CommitTransExit SubCFT_Abort:On Error Resume NextSet F = NothingSet td = NothingDBEngine(0).RollbackDBEngine(0).RollbackErr.ClearOn Error GoTo 0Err.Raise CFT_Failed, “ChangeFieldType”, E_DescExit SubCFT_Err:E_Desc = “Error ” & ProcessIf SubProcess <> “” Then E_Desc = E_Desc & vbCrLf & SubProcessIf DBEngine.Errors.Count = 0 ThenE_Desc = E_Desc & vbCrLf & “Error ” & Err.Number & ” ” & _Err.DescriptionElseFor Each E In DBEngine.ErrorsE_Desc = E_Desc & vbCrLf & “Error ” & E.Number & ” (” & _E.Source & “) ” & E.DescriptionNext EEnd IfDebug.Print E_DescResume CFT_AbortEnd SubPrivate Sub RecordRelationInfo(ByVal R As Relation, colR As Collection)’ Records information regarding the relationship and its fields’ in the colR collection.Dim F1 As Long, F As FieldFor F1 = 0 To R.Fields.Count – 1Set F = R.Fields(F1)colR.Add MakeArray(R.Name, R.Attributes, R.Table, R.ForeignTable, _F.Name, F.ForeignName)Next F1End SubPrivate Sub RecordIndexInfo(ByVal I As Index, colI As Collection)’ Records information about fields in the index and about the index itself’ into the colI collection.Dim F1 As Long, F As FieldFor F1 = 0 To I.Fields.Count – 1Set F = I.Fields(F1)colI.Add MakeArray(I.Name, I.Primary, I.Unique, I.Required, _I.IgnoreNulls, I.Clustered, F.Name, F.Attributes)Next F1End SubPrivate Function IsField(td As TableDef, ByVal FieldName As String) _As Boolean’ Returns TRUE if a field exists in the table with the same name as’specified in FieldName.’ Returns FALSE otherwise.Dim F As FieldErr.ClearOn Error Resume NextSet F = td(FieldName)IsField = Err.Number = 0Err.ClearEnd Function Private Function MakeArray(ParamArray X() As Variant) As Variant’ Does the same thing as the Array() function in VB6MakeArray = XEnd Function If necessary, change the CFT_Failed constant to use an error number that conforms to your company’s standards.Paste the following code into the General Declarations section of Form1’s Code Window:

Private Sub Command1_Click()Dim strDB As StringstrDB = “c:\Program Files\Microsoft Visual Studio\VB98\Nwind.mdb”Dim db As DAO.DatabaseSet db = DBEngine(0).OpenDatabase(strDB)ChangeFieldType db, “Customers”, “CustomerID”, dbText, 8db.CloseEnd Sub If necessary, modify strDB to use your Nwind database.Run the sample project.
Click the command button.
End the project.Examine the table in Microsoft Access or the Visual Basic Visual Database Manager add-in.
Note that the field has been resized.

Symptoms
In a Microsoft Visual Basic .NET collection, you can add a member after a zero index by using the Add method, but the collection is one-based. However, when you add a member after a zero index, the member is added at the first index.
Note In a Microsoft Visual Basic 6.0 collection, you receive a “Subscript out of range” error when you try to add a member after a zero index by using the Add method.
Resolution
This bug occurs because the collection is implemented as an array with Empty placeholder to adjust for 1 based array as the value for the zero index. This collection implementation allows you to add a member after the zero index.