How do I extract the data in a FoxPro memo field u

问题:

I'm writing a C# program to convert a FoxPro database to XML, and everything works except the memo field is blank. Is there something I'm missing to convert that bit?

I'm using C# .Net 3.5 SP1, Visual FoxPro 9 SP 1 OLE DB Driver. Connection string is okay, as all other data is being pulled properly.

When I converted the FoxPro database to SQL Server, the memo field is also blank there, so I can't convert twice.

回答1:

Ended up having to do some work myself, but maybe it can help someone else out in the future:

        public static object GetDbaseOrFoxproRawValue(string DBPath, string TableName, string ColumnName, 
        string CompareColumnName, string CompareValue, bool CompareColumnIsAutoKey)
    {
        using (BinaryReader read = new BinaryReader(File.Open(
            Path.Combine(DBPath, TableName + ".dbf"), FileMode.Open, FileAccess.Read, FileShare.ReadWrite)))
        {
            // Is it a type of file that I can handle?
            if (new byte[] { 0x02, 0x03, 0x30, 0x43, 0x63, 0x83, 0x8b,
                             0xcb, 0xf5, 0xfb }.Contains(read.ReadByte()))
            {
                // Skip date.
                read.BaseStream.Seek(3, SeekOrigin.Current);

                // Read useful datas...
                uint RecordCount = read.ReadUInt32();
                ushort FirstRecord = read.ReadUInt16();
                ushort RecordLength = read.ReadUInt16();
                int FieldCount = FirstRecord - 296 / 32;

                // Make sure things aren't stupid.
                ColumnName = ColumnName.ToLower();
                CompareColumnName = CompareColumnName.ToLower();

                // Find target column (field)
                string temp;
                UInt32 CompareFieldOffset = uint.MaxValue, FieldOffset = uint.MaxValue;
                byte CompareFieldLength = 0, FieldLength = 0;
                char FieldType = ' ';
                for (int i = 0; i < FieldCount; i++)
                {
                    read.BaseStream.Seek(32 + (i * 32), SeekOrigin.Begin);
                    temp = Encoding.ASCII.GetString(read.ReadBytes(11)).Replace("\0", "").ToLower();
                    if (temp == CompareColumnName)
                    {
                        read.ReadChar();
                        CompareFieldOffset = read.ReadUInt32();
                        CompareFieldLength = read.ReadByte();
                    }
                    if (temp == ColumnName)
                    {
                        FieldType = read.ReadChar();
                        FieldOffset = read.ReadUInt32();
                        FieldLength = read.ReadByte();
                    }

                    if (CompareFieldOffset != uint.MaxValue && FieldOffset != uint.MaxValue)
                        break;
                }

                // Make sure we can continue.
                if (CompareFieldOffset == uint.MaxValue || 
                    FieldOffset == uint.MaxValue) return null;

                // Iterate through each record to find the one we want.
                for (int index = 0; index < RecordCount; index++)
                {
                    read.BaseStream.Seek(FirstRecord + (index * RecordLength) + CompareFieldOffset, SeekOrigin.Begin);
                    temp = Encoding.Default.GetString(read.ReadBytes(CompareFieldLength)).Replace("\0", "");
                    if (temp == CompareValue)
                    {
                        read.BaseStream.Seek(FirstRecord + (index * RecordLength) + FieldOffset, SeekOrigin.Begin);
                        switch (FieldType)
                        {
                            case 'M':
                            case 'I': return read.ReadUInt32();
                            case 'C':
                            default: return Encoding.Default.GetString(read.ReadBytes(FieldLength)).Replace("\0", "");
                        }
                    }
                }
            }
            else
            {
                return null;
            }
        }

        return null;
    }

Just grab the result from that and use it as an index into the memo file (that code is pretty simple using the MSDN documentation).

回答2:

I am not terribly familiar with C# or FoxPro or SQL Server, so I cannot give you much advice in that regard.

However, if you cannot find a suitable driver, you may consider parsing the raw data and memo files yourself. Another question has dealt with this:

What's the easiest way to read a FoxPro DBF file from Python?

Believe it or not, these file formats are quite simple to parse should you decide to write your own C# parser. These specifications are available from Microsoft:

• Visual FoxPro File Structures
• Table File Structure
• Memo File Structure

回答3:

I use ODBC to link VFP 8 tables and the memo fields work with no problem. I don't know if OLEDB is different.

You might not have Visual FoxPro tables here. Many VFP systems use the same tables as the FoxPro 2 or dBase application they replaced. You could look at the file header or just try one of the other ODBC drivers to see if they work.