Click here to Skip to main content
15,867,308 members
Articles / Database Development / SQL Server

Reading Excel files using ODBC

Rate me:
Please Sign up or sign in to vote.
4.74/5 (19 votes)
12 Jan 2000CPOL 445.2K   5.5K   102   54
A discussion and demonstration of reading Excel files using ODBC
  • Download demo project - 20 Kb

    The Problem

    After contributing that article about writing into an Excel file I got tons of requests about how to read from one. Well, you asked for it...

    The main problem is that you can´t read an Excel file without previously having some formatting done. Microsoft refers to this in one of their KB papers. If somewhere out there finds there´s a way to do the reading whithout the formatting, please let me know...

    Another problem is the DSN you need to have installed in your ODBC Admin. This is not very useful because you don´t always know the name of the Excel file from the start.

    The last problem I´m dealing with here is generally doing ODBC reading using CRecordset without deriving from it. That is because if I always have to create a class for every single table I want to use, I´ll end up with lots of rather unnecessary code enlarging my app´s exe.

    The Solution

    1. According to Microsoft, an Excel sheet of version 4.x and later can only be read by ODBC if a database range is defined. Unfortunately they don´t state how to do this exactly. One way to let ODBC know what data is in there is to name a range of data on a worksheet using "Insert/Names" from Excel´s menu. There can be more than one "table" on a worksheet. This means that a sheet isn´t necessarily the same as a table in a "real" database. If you open "ReadExcel.xls" from the attached demo project and look up the names, you´ll see what I mean...
    2. Omiting the DSN tag in the connect string of CDatabase:Open() gives the opportunity to refer the ODBC-Driver directly using its name so we don´t have to have a DSN registered. This, of course, implies that the name of the ODBC-Driver is exactly known. If it isn´t, a call to SQLGetInstalledDrivers() will show all the installed drivers. For an example see CReadExcelDlg::GetExcelDriver() below.
    3. To use CRecordset the plain way you have to use a readonly, foreward only recordset. The data to get is defined by the SQL statement you put into CRecordset::Open(). Reading out the result is done by CRecordset::GetFieldValue(). For an example see the code below.

    What is needed

    In order to get the code below going you have to

  • include <afxdb.h>
  • include <odbcinst.h>
  • install an ODBC-driver called "MICROSOFT EXCEL DRIVER (*.XLS)" (or something like that)
  • You must use an ODBC Admin version 3.5 or higher

    Drawbacks

    Using a pseudo DSN only works with ODBC Admin V3.51 and higher. Earlier versions will not be able to use a DSN that actually isn´t installed. The result of an attempt to do so is some mumbling about missing registry keys.

    If using an underived CRecordset it needs to be readonly, foreward only. So any attempts to change the data or to move back will fail horribly. If you need to do something like that you´re bound to use CRecordset the "usual" way. Another drawback is that the tremendous overhead of CRecordset does in fact make it rather slow. A solution to this would be using the class CSQLDirect contributed by Dave Merner at http://www.codeguru.com/mfc_database/direct_sql_with_odbc.shtml.

    There´s still work to do

    One unsolved mystery in reading those files is how to get the data WITHOUT having a name defined for it. That means how can the structure of the data be retrieved, how many "tables" are in there, and so on. If you have any idea about that I´d be glad to read it under almikula@EUnet.at (please make a CC to alexander.mikula@siemens.at).

    The Source Code

    // Query an Excel file
    void CReadExcelDlg::OnButton1() 
    {
        CDatabase database;
        CString sSql;
        CString sItem1, sItem2;
        CString sDriver;
        CString sDsn;
        CString sFile = "ReadExcel.xls"; // the file name. Could also be something
                                         //  like C:\\Sheets\\WhatDoIKnow.xls
        
        // Clear the contents of the listbox
        m_ctrlList.ResetContent();
        
        // Retrieve the name of the Excel driver. This is 
        // necessary because Microsoft tends to use language
        // specific names like "Microsoft Excel Driver (*.xls)" versus
        // "Microsoft Excel Treiber (*.xls)"
        sDriver = GetExcelDriver();
        if (sDriver.IsEmpty())
        {
            // Blast! We didn´t find that driver!
            AfxMessageBox("No Excel ODBC driver found");
            return;
        }
        
        // Create a pseudo DSN including the name of the Driver and the Excel file
        // so we don´t have to have an explicit DSN installed in our ODBC admin
        sDsn.Format("ODBC;DRIVER={%s};DSN='';DBQ=%s", sDriver, sFile);
    
        TRY
        {
            // Open the database using the former created pseudo DSN
            database.Open(NULL, false, false, sDsn);
            
            // Allocate the recordset
            CRecordset recset(&database);
    
            // Build the SQL string
            // Remember to name a section of data in the Excel sheet using
            // "Insert->Names" to be able to work with the data like you would
            // with a table in a "real" database. There may be more than one table
            // contained in a worksheet.
            sSql = "SELECT field_1, field_2 "       
                   "FROM demo_table "                 
                   "ORDER BY field_1";
        
            // Execute that query (implicitly by opening the recordset)
            recset.Open(CRecordset::forwardOnly, sSql, CRecordset::readOnly);
    
            // Browse the result
            while (!recset.IsEOF())
            {
                // Read the result line
                recset.GetFieldValue("field_1", sItem1);
                recset.GetFieldValue("field_2", sItem2);
    
                // Insert result into the list
                m_ctrlList.AddString(sItem1 + " --> "+sItem2);
    
                // Skip to the next resultline
                recset.MoveNext();
            }
    
            // Close the database
            database.Close();
                                 
        }
        CATCH(CDBException, e)
        {
            // A database exception occured. Pop out the details...
            AfxMessageBox("Database error: " + e->m_strError);
        }
        END_CATCH;
    }
    
    
    // Get the name of the Excel-ODBC driver 
    // Contibuted by Christopher W. Backen - Thanx Christoper
    CString CReadExcelDlg::GetExcelDriver()
    {
        char szBuf[2001];
        WORD cbBufMax = 2000;
        WORD cbBufOut;
        char *pszBuf = szBuf;
        CString sDriver;
    
        // Get the names of the installed drivers
        // ("odbcinst.h" has to be included )
        if (!SQLGetInstalledDrivers(szBuf, cbBufMax, &cbBufOut))
            return "";
        
        // Search for the driver...
        do
        {
            if (strstr(pszBuf, "Excel") != 0)
            {
                // Found !
                sDriver = CString(pszBuf);
                break;
            }
            pszBuf = strchr(pszBuf, '\0') + 1;
        }
        while (pszBuf[1] != '\0');
    
        return sDriver;
    }

    Please refer the demo project (ReadExcelDlg.cpp) for more details.

  • License

    This article, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)


    Written By
    Austria Austria
    This member has not yet provided a Biography. Assume it's interesting and varied, and probably something to do with programming.

    Comments and Discussions

     
    Questionmissing mfc100.dll Pin
    yaoohmu7-Dec-11 21:47
    yaoohmu7-Dec-11 21:47 
    GeneralMy vote of 5 Pin
    Mahdi Nejadsahebi22-Oct-11 2:26
    Mahdi Nejadsahebi22-Oct-11 2:26 
    GeneralSize of field problem Pin
    yongdiego29-Jul-08 13:52
    yongdiego29-Jul-08 13:52 
    GeneralFiring query against EXCEL in Java servlet Pin
    Viral4-Dec-07 2:21
    Viral4-Dec-07 2:21 
    Hi,

    Please help me...

    I want to fire a query against EXCEL sheet in Java servlet.
    I have a EXCEL sheeet with ID & Name columns.
    In Java I am getting some Name from user input and want associated ID.

    Query I am using:
    String query = "Select * from [Sheet1$] where [Name] like "
    +"%"+search+"%";
    where search is Java variable where I am getting user input.

    It is giving me error
    "java.sql.SQLException: [Microsoft][ODBC Excel Driver] Syntax error in query expression '[Name] like %vir%'. "

    Please help me how can I use 'LIKE' operator for this query. Or any other workaround...

    Thanks...
    GeneralError in GetExcelDriver() Pin
    Jochen Arndt23-Aug-07 2:37
    professionalJochen Arndt23-Aug-07 2:37 
    GeneralWriting the files to ACCESS instead of Excel Pin
    zouris2-Jul-07 13:02
    zouris2-Jul-07 13:02 
    GeneralProblem running the example project Pin
    tantraseeker14-May-07 4:50
    tantraseeker14-May-07 4:50 
    GeneralRe: Problem running the example project Pin
    nitin_ap18-May-07 1:24
    nitin_ap18-May-07 1:24 
    GeneralReading from Excel Pin
    Biswajit Ghosh15-Mar-07 6:36
    Biswajit Ghosh15-Mar-07 6:36 
    GeneralI ha ve more than three rows , this program reads only 3 rows. Pin
    Rajendrappa5-Dec-06 20:44
    Rajendrappa5-Dec-06 20:44 
    GeneralRe: I ha ve more than three rows , this program reads only 3 rows. Pin
    Member 48216498-Jan-08 16:46
    Member 48216498-Jan-08 16:46 
    GeneralRe: I ha ve more than three rows , this program reads only 3 rows. Pin
    vakka25-Feb-09 23:49
    vakka25-Feb-09 23:49 
    Generalexcel project deployment Pin
    fatih isikhan11-Apr-06 3:59
    fatih isikhan11-Apr-06 3:59 
    GeneralReading from excel without formatting Pin
    hupu15-Dec-05 1:22
    hupu15-Dec-05 1:22 
    QuestionReading/Writing Excel sheets in unix using C++ Pin
    Tanz21-Sep-05 5:14
    Tanz21-Sep-05 5:14 
    AnswerRe: Reading/Writing Excel sheets in unix using C++ Pin
    logitecherrr22-Jul-06 3:06
    logitecherrr22-Jul-06 3:06 
    QuestionHow to use a SQL clause? Pin
    Disen30-Mar-05 20:05
    Disen30-Mar-05 20:05 
    GeneralDatabase Error Pin
    ektoplasma200021-Feb-05 1:21
    ektoplasma200021-Feb-05 1:21 
    QuestionHelp!! How to read 2 or more columns? Pin
    helen_kwan112-Jan-05 6:17
    helen_kwan112-Jan-05 6:17 
    AnswerRe: Help!! How to read 2 or more columns? Pin
    helen_kwan112-Jan-05 6:32
    helen_kwan112-Jan-05 6:32 
    AnswerRe: Help!! How to read 2 or more columns? Pin
    vakka25-Feb-09 23:57
    vakka25-Feb-09 23:57 
    GeneralAn Error message Pin
    oldfriend2-Dec-04 10:06
    oldfriend2-Dec-04 10:06 
    GeneralReading table names from excelfile. Pin
    ard-k2-Dec-04 1:28
    ard-k2-Dec-04 1:28 
    GeneralRe: Reading table names from excelfile. Pin
    JR Cooper13-Feb-06 9:25
    JR Cooper13-Feb-06 9:25 
    GeneralRe: Reading table names from excelfile. Pin
    markusschroth16-Jul-07 21:53
    markusschroth16-Jul-07 21:53 

    General General    News News    Suggestion Suggestion    Question Question    Bug Bug    Answer Answer    Joke Joke    Praise Praise    Rant Rant    Admin Admin   

    Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.