simplecsvparser.cpp

ATL开发指南PDF格式

// SimpleCSVParser.cpp: implementation of the CSimpleCSVParser class.
//
//////////////////////////////////////////////////////////////////////

#include "stdafx.h"
#include "SimpleCSVParser.h"
#include "SimpleCSVParser.h"

#ifdef _DEBUG
#undef THIS_FILE
static char THIS_FILE[]=__FILE__;
#define new DEBUG_NEW
#endif


//////////////////////////////////////////////////////////////////////
// Construction/Destruction
//////////////////////////////////////////////////////////////////////

CSimpleCSVParser::CSimpleCSVParser()
{
	m_pData = NULL;
	m_bInitialized = FALSE;
	m_pFirstRow = NULL;
	m_pLastRow = NULL;
	m_pCurRow = NULL;
	m_nRows = 0;
	m_nCols = 0;
	m_nCurRow = 0;
}

CSimpleCSVParser::~CSimpleCSVParser()
{
	Cleanup();
}

// This method opens the file passed and reads all
// of the data into memory. The file is expected
// to be a CSV file with the first record containing
// column names
BOOL CSimpleCSVParser::Initialize(LPCTSTR fileName)
{
	m_bInitialized = FALSE;
	m_pFirstRow = NULL;
	m_pLastRow = NULL;
	m_pCurRow = NULL;
	m_nRows = 0;
	m_nCols = 0;

	// Prepare for Initialize to be called more than once
	Cleanup();
	
	CFile file(fileName, CFile::modeRead | CFile::shareDenyNone);

	// allocate a buffer large enough to hold the file Data
	m_pData = new BYTE[file.GetLength() + 1];
	if (!m_pData)
		return FALSE;

	// Initialize the buffer with -1. -1 is what I decided to use
	// as a file delimiter. When I read the contents of the file
	// into the buffer it will overwrite all of the -1's but the 
	// last one.
	memset(m_pData, (unsigned char)-1, file.GetLength() + 1);

	// read the data into the buffer
	file.Read((void*)m_pData, file.GetLength());

	ParseColumns();

	InitializeRowPtrs();
	MoveFirst();

	m_bInitialized = TRUE;
	return TRUE;
}

void DeleteColumn(CColumnDefinition* pColDef)
{
	if (pColDef)
		delete pColDef;
}

void CSimpleCSVParser::Cleanup()
{
	if (m_pData)
	{
		delete [] m_pData;
		m_pData = NULL;
	}

	ForEachColumn(::DeleteColumn);
	m_columns.RemoveAll();
}

// Assumption: The first row of data contains the column information in
// the following format:
//		ColName:ColType[:ColSize(for string types)][:Precision(for decimal types)]
//
// Strings have the following format:
// ColName:ColType:ColSize
//
// Ints have the following format:
// ColName:ColType
//
// Floats have the following format:
// ColName:ColType:Precision
void CSimpleCSVParser::ParseColumns()
{
	CColumnDefinition*	pColDef;
	TCHAR				buf[256];
	TCHAR				colName[256];
	DWORD				colType = stringType;
	DWORD				colSize = 0;
	DWORD				colPrecision = 0;
	LPBYTE				pCurByte = m_pData;
	int					nStartByte = 0;
	int					nCurByte = 0;
	enum				States { getColName, getColType, getColSize, getColPrecision };
	States				nextState = getColName;

	// Only read until we reach the end of line
	while (pCurByte)
	{
		switch(*pCurByte)
		{
		// new column definition
		case '\r':
		case ',':
			{
				switch (nextState)
				{
				case getColType:
					memset(buf, NULL, sizeof(buf));
					memcpy(buf, &m_pData[nStartByte], nCurByte - nStartByte);
					colType = ColumnTypeFromString(buf);

					ASSERT((colType == dateType) || (colType == intType));
					nStartByte = nCurByte + 1; // skip the ','
					nextState = getColName;
					break;
				case getColSize:
					memset(buf, NULL, sizeof(buf));
					memcpy(buf, &m_pData[nStartByte], nCurByte - nStartByte);
					colSize = atol(buf);
					nStartByte = nCurByte + 1; // skip the ','
					nextState = getColName;
					break;
				case getColPrecision:
					memset(buf, NULL, sizeof(buf));
					memcpy(buf, &m_pData[nStartByte], nCurByte - nStartByte);
					colPrecision = atol(buf);
					nStartByte = nCurByte + 1; // skip the ','
					nextState = getColName;
					break;
				}
				pColDef = new CColumnDefinition();
				pColDef->SetColumnData(colName, colType, colSize, colPrecision);
				m_columns.Add((void*)pColDef);
				memset(colName, NULL, sizeof(colName));
				colType = stringType;
				colSize = 0;
				colPrecision = 0;
				break;
			}
		// column attribute
		case ':':
			{
				switch(nextState)
				{
				case getColName:
					memset(colName, NULL, sizeof(colName));
					memcpy(colName, &m_pData[nStartByte], nCurByte - nStartByte);
					nStartByte = nCurByte + 1; // skip the ':'
					nextState = getColType;
					break;
				case getColType:
					memset(buf, NULL, sizeof(buf));
					memcpy(buf, &m_pData[nStartByte], nCurByte - nStartByte);
					colType = ColumnTypeFromString(buf);

					ASSERT((colType == stringType) || (colType == floatType));
					nStartByte = nCurByte + 1; // skip the ':'
					if (colType == stringType)
						nextState = getColSize;
					else
						nextState = getColPrecision;
					break;
				case getColSize:
				case getColPrecision:
					ASSERT(FALSE);
					break;
				}
				break;
			}
		}
		
		// End of the line
		if (*pCurByte == '\r')
			break;

		nCurByte++;
		pCurByte++;
	}

	m_nCols = m_columns.GetSize();
	ASSERT(m_nCols > 0);
}

DWORD CSimpleCSVParser::ColumnTypeFromString(LPTSTR strColType)
{
	CString colType = strColType;

	if (colType == CString("integer"))
		return intType;
	if (colType == CString("float"))
		return floatType;
	if (colType == CString("date"))
		return dateType;
	
	// by default assume string data type
	return stringType;
}


void DisplayColumn(CColumnDefinition* pColDef)
{
	if (pColDef)
		AfxMessageBox(pColDef->GetColumnName());
}

void CSimpleCSVParser::DisplayColumns()
{
	ForEachColumn(::DisplayColumn);
}

void CSimpleCSVParser::InitializeRowPtrs()
{
	m_pFirstRow = NULL;
	m_pLastRow = NULL;
	m_pCurRow = NULL;

	// The First Row will be the 2nd row in the file since
	// the first row contains column definitions
	m_pFirstRow = FindNextRow(m_pData);
	m_pCurRow = m_pFirstRow;

	LPBYTE pHoldPos = m_pFirstRow;
	m_nRows = 0;
	while (pHoldPos != NULL)
	{
		m_pLastRow = pHoldPos;
		pHoldPos = FindNextRow(pHoldPos);
		m_nRows++;
	}

}

LPBYTE CSimpleCSVParser::FindPrevRow(LPBYTE pPos)
{
	ASSERT(pPos);
	if (!pPos)
		return NULL;

	ASSERT(m_pFirstRow != NULL);
	if (m_pFirstRow == NULL)
		return NULL;

	if (pPos == m_pFirstRow)
		return m_pFirstRow;

	BOOL bInPrevRow = FALSE;
	LPBYTE p = pPos;
	while (p != m_pFirstRow)
	{
		if (*p == '\r' || *p == '\n')
		{
			if (bInPrevRow)
			{
				p++;
				break;
			}
			else
			{
				while (*p == '\r' || *p == '\n')
				{
					p--;
				}
				bInPrevRow = TRUE;
			}
		}
		p--;
	}

	return p;
}

LPBYTE CSimpleCSVParser::FindNextRow(LPBYTE pPos)
{
	ASSERT(pPos);
	if (!pPos)
		return NULL;

	LPBYTE p = pPos;
	unsigned char endOfFile = (unsigned char)-1;
	
	while (*p != endOfFile)
	{
		if (*p == '\r' || *p == '\n')
		{
			while (*p == '\r' || *p == '\n')
			{
				p++;
			}

			return ((*p != endOfFile) ? p : NULL);
		}
		p++;
	}

	return NULL;
}

void CSimpleCSVParser::ForEachColumn(ForEachColumnFunc pForEachColumnFunc)
{
	if (!m_bInitialized)
		return;

	for (int i = 0; i < m_columns.GetSize(); i++)
	{
		CColumnDefinition* pColDef = (CColumnDefinition*) m_columns[i];
		(*pForEachColumnFunc)(pColDef);
	}
}

BOOL CSimpleCSVParser::ParseCurrentRow()
{
	CleanColumnData();

	ASSERT(m_pCurRow != NULL);
	if (m_pCurRow == NULL)
		return FALSE;

	BOOL bEndOfRow = FALSE;
	CString colValue;
	LPBYTE pStartPos = m_pCurRow;
	LPBYTE pEndPos = m_pCurRow;
	int curCol = 0;

	while (pEndPos)
	{
		switch(*pEndPos)
		{
		// new column definition
		case '\r':
		case 255: // end of file
			bEndOfRow = TRUE;
			colValue.Empty();
			if (pEndPos > pStartPos)
			{
				LPTSTR pBuffer = colValue.GetBuffer(pEndPos - pStartPos);
				_tcsncpy(pBuffer, (const char*)pStartPos, pEndPos - pStartPos);
				colValue.ReleaseBuffer();
				((CColumnDefinition*)m_columns.GetAt(curCol))->SetColumnValue(colValue);
			}

			curCol++;
			break;
		case ',':
			colValue.Empty();
			if (pEndPos > pStartPos)
			{
				LPTSTR pBuffer = colValue.GetBuffer(pEndPos - pStartPos);
				_tcsncpy(pBuffer, (const char*)pStartPos, pEndPos - pStartPos);
				colValue.ReleaseBuffer();
				((CColumnDefinition*)m_columns.GetAt(curCol))->SetColumnValue(colValue);
			}

			curCol++;
			pEndPos++;
			pStartPos = pEndPos;
			break;
		}

		if (bEndOfRow)
			break;

		pEndPos++;
	}
	return TRUE;
}

void EmptyData(CColumnDefinition* pColDef)
{
	if (pColDef)
		pColDef->SetColumnValue(_T(""));
}

void CSimpleCSVParser::CleanColumnData()
{
	ForEachColumn(EmptyData);
}

BOOL CSimpleCSVParser::MoveFirst()
{
	ASSERT(m_pFirstRow);
	if (!m_pFirstRow)
		return FALSE;

	m_nCurRow = 1;
	m_pCurRow = m_pFirstRow;
	return ParseCurrentRow();
}

BOOL CSimpleCSVParser::MoveNext()
{
	ASSERT(m_pCurRow);
	if (!m_pCurRow)
		return FALSE;

	ASSERT(m_pLastRow);
	if (!m_pLastRow)
		return FALSE;

	if (m_pCurRow == m_pLastRow) 
		return FALSE;

	m_pCurRow = FindNextRow(m_pCurRow);
	if (ParseCurrentRow())
		m_nCurRow++;
	else
		return FALSE;
	return TRUE;
}

BOOL CSimpleCSVParser::MovePrev()
{
	ASSERT(m_pCurRow);
	if (!m_pCurRow)
		return FALSE;

	ASSERT(m_pFirstRow);
	if (!m_pFirstRow)
		return FALSE;

	if (m_pCurRow == m_pFirstRow) 
		return FALSE;

	m_pCurRow = FindPrevRow(m_pCurRow);
	if (ParseCurrentRow())
		m_nCurRow--;
	else
		return FALSE;
	return TRUE;
}

BOOL CSimpleCSVParser::MoveLast()
{
	ASSERT(m_pLastRow);
	if (!m_pLastRow)
		return FALSE;

	m_pCurRow = m_pLastRow;
	if (ParseCurrentRow())
		m_nCurRow = m_nRows;
	else
		return FALSE;
	return TRUE;
}

long CSimpleCSVParser::GetColumnCount()
{
	return (long)m_columns.GetSize();
}

long CSimpleCSVParser::GetRowCount() const
{
	return (long)m_nRows;
}

CColumnDefinition* CSimpleCSVParser::GetColumnDefinition(int ndx)
{
	return (CColumnDefinition*)m_columns[ndx];
}

BOOL CSimpleCSVParser::GotoRow(long nIndex)
{
	if (nIndex < 0)
		return FALSE;
	if (nIndex >= m_nRows)
		return FALSE;

	if (nIndex == 0)
		MoveFirst();
	else if (nIndex == (m_nRows - 1))
		MoveLast();
	else if (nIndex > m_nCurRow)
	{
		while (nIndex > m_nCurRow)
		{
			MoveNext();
		}
	}
	else if (nIndex < m_nCurRow)
	{
		while (nIndex < m_nCurRow)
		{
			MovePrev();
		}
	}

	return TRUE;
}