我必须将.DBF和.FPT文件从Visual FoxPro转换为MySQL。现在我的脚本适用于.DBF文件，它打开并用dbase_open（）和dbase_get_record_with_names（）读取它们，然后执行MySQL INSERT命令。
但是，这些.DBF文件的某些字段属于MEMO类型，因此存储在以.FPT结尾的单独文件中。我该如何阅读此文件？
我在MSDN中找到了这个文件类型的规范，但我不知道如何用PHP按字节顺序读取这个文件（另外，我更喜欢简化的解决方案）。
有什么想法吗？
答案 0 :(得分：12)
好的，我仔细研究了DBF和FPT文件结构的MSDN规范，结果是一个漂亮的PHP类，可以同时打开DBF和（可选）FPT备忘录文件。此类将为您提供记录后的记录，从而从备忘录文件中获取任何备忘录 - 如果已打开。
class Prodigy_DBF {
private $Filename, $DB_Type, $DB_Update, $DB_Records, $DB_FirstData, $DB_RecordLength, $DB_Flags, $DB_CodePageMark, $DB_Fields, $FileHandle, $FileOpened;
private $Memo_Handle, $Memo_Opened, $Memo_BlockSize;
private function Initialize() {
if($this->FileOpened) {
fclose($this->FileHandle);
}
if($this->Memo_Opened) {
fclose($this->Memo_Handle);
}
$this->FileOpened = false;
$this->FileHandle = NULL;
$this->Filename = NULL;
$this->DB_Type = NULL;
$this->DB_Update = NULL;
$this->DB_Records = NULL;
$this->DB_FirstData = NULL;
$this->DB_RecordLength = NULL;
$this->DB_CodePageMark = NULL;
$this->DB_Flags = NULL;
$this->DB_Fields = array();
$this->Memo_Handle = NULL;
$this->Memo_Opened = false;
$this->Memo_BlockSize = NULL;
}
public function __construct($Filename, $MemoFilename = NULL) {
$this->Prodigy_DBF($Filename, $MemoFilename);
}
public function Prodigy_DBF($Filename, $MemoFilename = NULL) {
$this->Initialize();
$this->OpenDatabase($Filename, $MemoFilename);
}
public function OpenDatabase($Filename, $MemoFilename = NULL) {
$Return = false;
$this->Initialize();
$this->FileHandle = fopen($Filename, "r");
if($this->FileHandle) {
// DB Open, reading headers
$this->DB_Type = dechex(ord(fread($this->FileHandle, 1)));
$LUPD = fread($this->FileHandle, 3);
$this->DB_Update = ord($LUPD[0])."/".ord($LUPD[1])."/".ord($LUPD[2]);
$Rec = unpack("V", fread($this->FileHandle, 4));
$this->DB_Records = $Rec[1];
$Pos = fread($this->FileHandle, 2);
$this->DB_FirstData = (ord($Pos[0]) + ord($Pos[1]) * 256);
$Len = fread($this->FileHandle, 2);
$this->DB_RecordLength = (ord($Len[0]) + ord($Len[1]) * 256);
fseek($this->FileHandle, 28); // Ignoring "reserved" bytes, jumping to table flags
$this->DB_Flags = dechex(ord(fread($this->FileHandle, 1)));
$this->DB_CodePageMark = ord(fread($this->FileHandle, 1));
fseek($this->FileHandle, 2, SEEK_CUR); // Ignoring next 2 "reserved" bytes
// Now reading field captions and attributes
while(!feof($this->FileHandle)) {
// Checking for end of header
if(ord(fread($this->FileHandle, 1)) == 13) {
break; // End of header!
} else {
// Go back
fseek($this->FileHandle, -1, SEEK_CUR);
}
$Field["Name"] = trim(fread($this->FileHandle, 11));
$Field["Type"] = fread($this->FileHandle, 1);
fseek($this->FileHandle, 4, SEEK_CUR); // Skipping attribute "displacement"
$Field["Size"] = ord(fread($this->FileHandle, 1));
fseek($this->FileHandle, 15, SEEK_CUR); // Skipping any remaining attributes
$this->DB_Fields[] = $Field;
}
// Setting file pointer to the first record
fseek($this->FileHandle, $this->DB_FirstData);
$this->FileOpened = true;
// Open memo file, if exists
if(!empty($MemoFilename) and file_exists($MemoFilename) and preg_match("%^(.+).fpt$%i", $MemoFilename)) {
$this->Memo_Handle = fopen($MemoFilename, "r");
if($this->Memo_Handle) {
$this->Memo_Opened = true;
// Getting block size
fseek($this->Memo_Handle, 6);
$Data = unpack("n", fread($this->Memo_Handle, 2));
$this->Memo_BlockSize = $Data[1];
}
}
}
return $Return;
}
public function GetNextRecord($FieldCaptions = false) {
$Return = NULL;
$Record = array();
if(!$this->FileOpened) {
$Return = false;
} elseif(feof($this->FileHandle)) {
$Return = NULL;
} else {
// File open and not EOF
fseek($this->FileHandle, 1, SEEK_CUR); // Ignoring DELETE flag
foreach($this->DB_Fields as $Field) {
$RawData = fread($this->FileHandle, $Field["Size"]);
// Checking for memo reference
if($Field["Type"] == "M" and $Field["Size"] == 4 and !empty($RawData)) {
// Binary Memo reference
$Memo_BO = unpack("V", $RawData);
if($this->Memo_Opened and $Memo_BO != 0) {
fseek($this->Memo_Handle, $Memo_BO[1] * $this->Memo_BlockSize);
$Type = unpack("N", fread($this->Memo_Handle, 4));
if($Type[1] == "1") {
$Len = unpack("N", fread($this->Memo_Handle, 4));
$Value = trim(fread($this->Memo_Handle, $Len[1]));
} else {
// Pictures will not be shown
$Value = "{BINARY_PICTURE}";
}
} else {
$Value = "{NO_MEMO_FILE_OPEN}";
}
} else {
$Value = trim($RawData);
}
if($FieldCaptions) {
$Record[$Field["Name"]] = $Value;
} else {
$Record[] = $Value;
}
}
$Return = $Record;
}
return $Return;
}
function __destruct() {
// Cleanly close any open files before destruction
$this->Initialize();
}
}
这个类可以像这样使用：
$Test = new Prodigy_DBF("customer.DBF", "customer.FPT");
while(($Record = $Test->GetNextRecord(true)) and !empty($Record)) {
print_r($Record);
}
它可能不是一个全能的完美课程，但它对我有用。请随意使用此代码，但请注意，该类非常宽容 - 它不关心fread（）和fseek（）是否返回true或其他任何东西 - 所以您可能希望在使用之前稍微改进它。
另请注意，有许多私人变量，例如记录数，记录大小等，目前尚未使用。
答案 1 :(得分：3)
我替换了这段代码：
fseek($this->Memo_Handle, (trim($RawData) * $this->Memo_BlockSize)+8);
$Value = trim(fread($this->Memo_Handle, $this->Memo_BlockSize));
使用此代码：
fseek($this->Memo_Handle, (trim($RawData) * $this->Memo_BlockSize)+4);
$Len = unpack("N", fread($this->Memo_Handle, 4));
$Value = trim(fread($this->Memo_Handle, $Len[1] ));
这对我有帮助
答案 2 :(得分：1)
虽然不是PHP，但VFP是基于1的引用，我认为PHP是基于零的引用，所以你必须对decypher进行相应的调整，但是这样可行，希望你能够 完成后发布此部分的版本。
VFP中的FILETOSTR（）将打开一个文件，并将整个内容读入 单个内存变量作为字符串 - 所有转义键，高字节字符等都完好无损。您可能需要依赖FOPEN（），FSEEK（），FCLOSE（）等。
MemoTest.FPT是我的示例备忘录表/文件 fpt1 = FILETOSTR（“MEMOTEST.FPT”）
首先，您必须检测创建文件时使用的MEMO BLOCK SIZE。通常，这将是64 BYTES，但是根据您在帖子中的链接。
标题位置6-7标识大小（VFP，位置7和8）。第一个字节是高阶
nBlockSize = ASC( SUBSTR( fpt1, 7, 1 )) * 256 + ASC( SUBSTR( fpt1, 8, 1 ))
现在，根据您的个人记录。 DBF结构中的任何地方都有备忘录FIELD（并且每个记录结构可以有很多），将有4个字节。在“记录”字段中，它标识备忘录文件中存储内容的“块”。
MemoBytes =您指定的字段位置的4个字节。这些将以0-255的ASCII格式存储。该字段与FIRST字节一起存储为低位，第四字节存储为256 ^ 3 = 16777216.第一个“块”将使用头文件占用的备忘录.fpt文件规范的位置偏移量512开始职位0-511。
因此，如果你的第一个备忘录字段的内容为“8000”，其中8是实际的0x08，而不是数字“8”，即0x38，零是0x00。
YourMemoField =“8000”（实际使用ascii，但为了显示十六进制预期值的可读性）
First Byte is ASCII value * 1 ( 256 ^ 0 )
Second Byte is ASCII value * 256 (256 ^ 1)
Third Byte is ASCII value * 65536 (256 ^ 2)
Fourth Byte is ASCII value * 16777216 (256 ^ 3)
nMemoBlock = byte1 + ( byte2 * 256 ) + ( byte3 * 65536 ) + ( byte4 * 16777216 )
现在，你需要FSEEK（）到
FSEEK( handle, nMemoBlock * nBlockSize +1 )
表示您要查找的块的第一个字节。这将指向BLOCK标头。在这种情况下，根据规范，前4个字节标识块SIGNATURE，后4个字节是内容的长度。对于这两个，字节首先用HIGH-BYTE存储。
从你的FSEEK（），它的上面的nMemoBlock的REVERSE与高字节。这里的“Byte1-4”来自您的FSEEK（）位置
nSignature = ( byte1 * 16777216 ) + ( byte2 * 65536 ) + ( byte3 * 256 ) + byte4
nMemoLength = ( byte5 * 16777216 ) + ( byte6 * 65536 ) + ( byte7 * 256 ) + byte8
现在，FSEEK（）到第9个字节（数据的第一个实际字符，你刚刚读取的标题的8个字节用于签名和备忘录长度）。这是数据的开始。
现在，阅读其余内容......
FSEEK() +9 characters to new position
cFinalMemoData = FREAD( handle, nMemoLength )
我知道这不是完美的，也不是PHP脚本，但它足够的伪代码存储在什么东西上，希望能让你顺利上路。
同样，请仔细考虑，因为您正在逐步完成调试过程以确保0或1偏移量。为了帮助简化和测试，我创建了一个简单的.DBF，包含2个字段......一个字符字段和一个备忘录字段，添加了一些记录和一些基本内容来确认所有内容，位置等。
答案 3 :(得分：1)
<?
class Prodigy_DBF {
private $Filename, $DB_Type, $DB_Update, $DB_Records, $DB_FirstData, $DB_RecordLength, $DB_Flags, $DB_CodePageMark, $DB_Fields, $FileHandle, $FileOpened;
private $Memo_Handle, $Memo_Opened, $Memo_BlockSize;
private function Initialize() {
if($this->FileOpened) {
fclose($this->FileHandle);
}
if($this->Memo_Opened) {
fclose($this->Memo_Handle);
}
$this->FileOpened = false;
$this->FileHandle = NULL;
$this->Filename = NULL;
$this->DB_Type = NULL;
$this->DB_Update = NULL;
$this->DB_Records = NULL;
$this->DB_FirstData = NULL;
$this->DB_RecordLength = NULL;
$this->DB_CodePageMark = NULL;
$this->DB_Flags = NULL;
$this->DB_Fields = array();
$this->Memo_Handle = NULL;
$this->Memo_Opened = false;
$this->Memo_BlockSize = NULL;
}
public function __construct($Filename, $MemoFilename = NULL) {
$this->Prodigy_DBF($Filename, $MemoFilename);
}
public function Prodigy_DBF($Filename, $MemoFilename = NULL) {
$this->Initialize();
$this->OpenDatabase($Filename, $MemoFilename);
}
public function OpenDatabase($Filename, $MemoFilename = NULL) {
$Return = false;
$this->Initialize();
$this->FileHandle = fopen($Filename, "r");
if($this->FileHandle) {
// DB Open, reading headers
$this->DB_Type = dechex(ord(fread($this->FileHandle, 1)));
$LUPD = fread($this->FileHandle, 3);
$this->DB_Update = ord($LUPD[0])."/".ord($LUPD[1])."/".ord($LUPD[2]);
$Rec = unpack("V", fread($this->FileHandle, 4));
$this->DB_Records = $Rec[1];
$Pos = fread($this->FileHandle, 2);
$this->DB_FirstData = (ord($Pos[0]) + ord($Pos[1]) * 256);
$Len = fread($this->FileHandle, 2);
$this->DB_RecordLength = (ord($Len[0]) + ord($Len[1]) * 256);
fseek($this->FileHandle, 28); // Ignoring "reserved" bytes, jumping to table flags
$this->DB_Flags = dechex(ord(fread($this->FileHandle, 1)));
$this->DB_CodePageMark = ord(fread($this->FileHandle, 1));
fseek($this->FileHandle, 2, SEEK_CUR); // Ignoring next 2 "reserved" bytes
// Now reading field captions and attributes
while(!feof($this->FileHandle)) {
// Checking for end of header
if(ord(fread($this->FileHandle, 1)) == 13) {
break; // End of header!
} else {
// Go back
fseek($this->FileHandle, -1, SEEK_CUR);
}
$Field["Name"] = trim(fread($this->FileHandle, 11));
$Field["Type"] = fread($this->FileHandle, 1);
fseek($this->FileHandle, 4, SEEK_CUR); // Skipping attribute "displacement"
$Field["Size"] = ord(fread($this->FileHandle, 1));
fseek($this->FileHandle, 15, SEEK_CUR); // Skipping any remaining attributes
$this->DB_Fields[] = $Field;
}
// Setting file pointer to the first record
fseek($this->FileHandle, $this->DB_FirstData);
$this->FileOpened = true;
// Open memo file, if exists
if(!empty($MemoFilename) and file_exists($MemoFilename) and preg_match("%^(.+).fpt$%i", $MemoFilename)) {
$this->Memo_Handle = fopen($MemoFilename, "r");
if($this->Memo_Handle) {
$this->Memo_Opened = true;
// Getting block size
fseek($this->Memo_Handle, 6);
$Data = unpack("n", fread($this->Memo_Handle, 2));
$this->Memo_BlockSize = $Data[1];
}
}
}
return $Return;
}
public function GetNextRecord($FieldCaptions = false) {
$Return = NULL;
$Record = array();
if(!$this->FileOpened) {
$Return = false;
} elseif(feof($this->FileHandle)) {
$Return = NULL;
} else {
// File open and not EOF
fseek($this->FileHandle, 1, SEEK_CUR); // Ignoring DELETE flag
foreach($this->DB_Fields as $Field) {
$RawData = fread($this->FileHandle, $Field["Size"]);
// Checking for memo reference
if($Field["Type"] == "M" and $Field["Size"] == 4 and !empty($RawData)) {
// Binary Memo reference
$Memo_BO = unpack("V", $RawData);
if($this->Memo_Opened and $Memo_BO != 0) {
fseek($this->Memo_Handle, $Memo_BO[1] * $this->Memo_BlockSize);
$Type = unpack("N", fread($this->Memo_Handle, 4));
if($Type[1] == "1") {
$Len = unpack("N", fread($this->Memo_Handle, 4));
$Value = trim(fread($this->Memo_Handle, $Len[1]));
} else {
// Pictures will not be shown
$Value = "{BINARY_PICTURE}";
}
} else {
$Value = "{NO_MEMO_FILE_OPEN}";
}
} else {
if($Field["Type"] == "M"){
if(trim($RawData) > 0) {
fseek($this->Memo_Handle, (trim($RawData) * $this->Memo_BlockSize)+8);
$Value = trim(fread($this->Memo_Handle, $this->Memo_BlockSize));
}
}else{
$Value = trim($RawData);
}
}
if($FieldCaptions) {
$Record[$Field["Name"]] = $Value;
} else {
$Record[] = $Value;
}
}
$Return = $Record;
}
return $Return;
}
function __destruct() {
// Cleanly close any open files before destruction
$this->Initialize();
}
}
?>
答案 4 :(得分：0)
我认为在PHP中不太可能有FoxPro库。
您可能需要从头开始编码。对于按字节读取，请遇到fopen() fread() and colleagues。
编辑：似乎有一个Visual FoxPro ODBC driver。您可以通过PDO和该连接器连接到FoxPro数据库。我不知道成功的机会如何，以及它将会有多少工作。
答案 5 :(得分：0)
FPT文件包含备忘录数据。在DBF中，您有类型为memo的列，此列中的信息是指向FPT文件中条目的指针。
如果要查询表中的数据，则只需引用备注列即可获取数据。您不需要单独解析FPT文件中的数据。 OLE DB驱动程序（如果您的文件是VFP 6或更早版本，则为ODBC驱动程序）应该只提供此信息。
有一个工具可以自动将Visual FoxPro数据迁移到MySQL。你可能想看看它是否可以节省一些时间：
并搜索“Stru2MySQL_2”以获取用于迁移数据结构的工具，并搜索“VFP2MySQL数据上载程序”以获取有助于迁移的工具。
Rick Schummer VFP MVP
答案 6 :(得分：0)
您可能还想查看PHP dbase库。它们可以很好地处理DBF文件。