读取文件然后将文件写入 SQL Server BLOB 列的代码，生成具有不同字节的文件

Posted 2023-04-18

【中文标题】读取文件然后将文件写入 SQL Server BLOB 列的代码，生成具有不同字节的文件

【英文标题】：Code to read and then write a file into a SQL Server BLOB column producing files with different bytes

【发布时间】：2017-01-29 11:34:12

【问题描述】：

使用：

MS-SQL Server 2014 MS-Access 2013 带有链接到 SQL 的 ODBC 表 服务器数据库 “SQL Server”ODBC 驱动程序（10.00.10586.00，Microsoft 公司，SQLSRV32.DLL) 道

我们有一个 Access 数据库，其中将 ODBC 表链接到 SQL Server 2014 数据库，在 Access 应用程序中的表单后面使用 VBA 代码将文件上传到 SQL Server blob (varbinary[max]) 列，然后下载来自同一 blob 列的文件。

但是我们发现，在检索之前从 blob 列上传的文件时，保存的文件在文件末尾添加了一些额外的字节。

Beyond Compare 中两个文件的比较屏幕截图如下：

如果有人可以检查并指出代码中的错误，我将不胜感激。代码如下：

Function ReadBLOB(SourceFileName As String, TableName As String, FieldName As String, _
                  IDFieldName As String, IDFieldValue As Variant)
    Dim NumBlocks As Integer, SourceFile As Integer, i As Integer
    Dim FileLength As Long
    Dim LeftOver As Long
    Dim FileData() As Byte
    Dim RetVal As Variant
    Dim BlockSize As Long

    Dim s As String

    On Error GoTo Err_ReadBLOB

    BlockSize = 32767

    ' Open the source file.
    SourceFile = FreeFile
    Open SourceFileName For Binary Access Read As SourceFile

    ' Get the length of the file.
    FileLength = LOF(SourceFile)
    If FileLength = 0 Then
        ReadBLOB = 0
        Exit Function
    End If

    ' Calculate the number of blocks to read and leftover bytes.
    NumBlocks = FileLength \ BlockSize
    LeftOver = FileLength Mod BlockSize

    Dim T As dao.Recordset

    If TypeName(IDFieldValue) = "String" Then
        IDFieldValue = "'" & IDFieldValue & "'"
    End If

    s = "SELECT [" & FieldName & "] FROM [" & TableName & "] WHERE [" & IDFieldName & "] = " & IDFieldValue

    Set T = CurrentDb.OpenRecordset(s, dbOpenDynaset, dbSeeChanges)

    T.Edit

    ' Read the 1st block of data (upto Leftover in size), writing it to the table.
    'FileData = String$(LeftOver, 32)
    ReDim FileData(LeftOver)
    Get SourceFile, , FileData
    T(FieldName).AppendChunk (FileData)

    ' Read the remaining blocks of data, writing them to the table.
    'FileData = String$(BlockSize, 32)
    ReDim FileData(BlockSize)
    For i = 1 To NumBlocks
        Get SourceFile, , FileData
        T(FieldName).AppendChunk (FileData)

    Next i

    ' Update the record and terminate function.
    T.Update
    Close SourceFile
    ReadBLOB = FileLength
    Exit Function

Err_ReadBLOB:
    ReadBLOB = -Err

    MsgBox Err.Description

    Exit Function
End Function

Function WriteBLOB2(TableName As String, FieldName As String, IDFieldName As String, _
                    IDFieldValue As Variant, DestinationFileName As String) As Long

    Dim NumBlocks As Integer, DestFile As Integer, i As Integer
    Dim FileLength As Long, LeftOver As Long
    Dim FileData() As Byte
    Dim RetVal As Variant
    Dim BlockSize As Long
    Dim s As String
    Dim f As String

    On Error GoTo Err_WriteBLOB

    BlockSize = 32767

    Dim T As dao.Recordset

    If TypeName(IDFieldValue) = "String" Then
        IDFieldValue = "'" & IDFieldValue & "'"
    End If

    s = "SELECT [" & FieldName & "] FROM [" & TableName & "] WHERE [" & IDFieldName & "] = " & IDFieldValue

    Set T = CurrentDb.OpenRecordset(s, dbOpenSnapshot, dbSeeChanges)

    If T.RecordCount = 0 Then
        WriteBLOB2 = 0
        Exit Function
    End If

    ' Get the size of the field.
    FileLength = T(FieldName).FieldSize()
    If FileLength = 0 Then
        WriteBLOB2 = 0
        Exit Function
    End If

    ' Calculate number of blocks to write and leftover bytes.
    NumBlocks = FileLength \ BlockSize
    LeftOver = FileLength Mod BlockSize

    ' Remove any existing destination file.
    DestFile = FreeFile
    Open DestinationFileName For Output As DestFile
    Close DestFile

    ' Open the destination file.
    Open DestinationFileName For Binary As DestFile

    ' Write the leftover data to the output file.
    FileData = T(FieldName).GetChunk(0, LeftOver)
    Put DestFile, , FileData

    ' Write the remaining blocks of data to the output file.
    For i = 1 To NumBlocks
        ' Reads a chunk and writes it to output file.
        FileData = T(FieldName).GetChunk((i - 1) * BlockSize + LeftOver, BlockSize)
        Put DestFile, , FileData

    Next i

    ' Terminates function
    Close DestFile
    WriteBLOB2 = FileLength
    Exit Function

Err_WriteBLOB:
    WriteBLOB2 = -Err

    MsgBox Err.Description

    Exit Function
End Function

Public Sub ClearSQLBlob2(TableName As String, FieldName As String, _
                         IDFieldName As String, IDFieldValue As Variant)

    If TypeName(IDFieldValue) = "String" Then
        IDFieldValue = "'" & IDFieldValue & "'"
    End If

    DoCmd.SetWarnings False
    DoCmd.RunSQL "UPDATE [" & TableName & "] SET [" & FieldName & "] = NULL WHERE [" & IDFieldName & "] = " & IDFieldValue
    DoCmd.SetWarnings True
End Sub

【问题讨论】：

这将有助于从您的问题中删除所有非必要代码，尤其是所有那些SysCmd 电话。

您是否 100% 确定 blob 列中的值包含原始数据？还是在上传过程中数据已经损坏？最后附加了哪些值？

@erg：文件在上传过程中损坏。我通过使用 blob 编辑器程序从数据库下载原始文件进行了检查。但这就是我问这个问题的原因。

【参考方案1】：

我认为问题是：

除非您的模块中有 Option Base 1 声明，否则数组是从零开始的。

所以如果例如LeftOver = 2,

ReDim FileData(LeftOver)

实际上会声明一个数组FileData(0 To 2)，它包含3个字节。所以下面的Get 将读取 3 个字节，但您希望它读取 2 个字节。

全尺寸数组也是如此。

最后你从文件中读取了太多NumBlocks + 1字节，剩下的将是00字节。

解决方案：使用

ReDim FileData(1 To LeftOver)
ReDim FileData(1 To BlockSize)

编辑：请注意，您必须检查LeftOver = 0 的情况。