[!IMPORTANT] Unfortunately I don't have the time to give this project the love it deserves. The source is "as-is" and no issues will be addressed currently. The docs have been imported from the original trac-wiki and only some parts have been transformed to proper markdown.
exOdbc is an open source C++ library to access ODBC data sources. It was inspired by wxOdbc, but has been rewritten completely. At its core exOdbc uses a boost::variant to store the different types of column buffers.
See TestedDatabases for an overview of Databases and ODBC drivers exOdbc has been tested with.
exOdbc is released under the GNU LESSER GENERAL PUBLIC LICENSE, see exOdbcLicense.
exOdbc uses CMake to support different plattforms. It has been tested on Windows and Linux.
Last Release is https://github.com/eg-be/exodbc/releases/tag/exOdbc_0.9.4
Make sure you have CMake installed and that you have a copy of the boost library ready. To build the tests, you need compiled libraries of boost-filesystem and boost-system.
After extracting the tarball, cd into the top-level directory and:
mkdir build
cd build
# On Winodws, indicate the path to the boost libraries:
set BOOST_ROOT=e:/lib/boost_1_62_0
# Build only the library and the samples:
cmake -DBUILD_TESTS=OFF ..
cmake --build .
Resulting binaries have been placed in bin
.
See INSTALL.txt for more cmake configuration options.
exOdbc uses ggogletest to test the functionality against different databases and ODBC drivers.
To build the tests, set cmake option BUILD_TESTS
to ON
(default).
mkdir build
cd build
# On Winodws, indicate the path to the boost libraries:
# Make sure that compiled version of boost-system and boost-filesystem are ready.
set BOOST_ROOT=e:/lib/boost_1_62_0
# Build everything:
cmake ..
cmake --build .
Resulting test executable exodbctest
has been placed in bin
.
Most of the tests require a database with some tables. See exOdbcTest for how to prepare the test database.
TestedDatabases has more information about the tested databases and the known failures of exOdbc.
Samples are available online with some additional Description.
The following sample demonstrates basic usage of exOdbc:
See samples/ShortIntro for updated versions.
/*!
* \file ShortIntro.cpp
* \author Elias Gerber <eg@elisium.ch>
* \date 03.01.2016
* \copyright GNU Lesser General Public License Version 3
*
* ShortIntro Sample.
*/
#ifdef _WIN32
#include <SDKDDKVer.h>
#include <tchar.h>
#endif
#include <iostream>
#include "exodbc/exOdbc.h"
#include "exodbc/Environment.h"
#include "exodbc/Database.h"
#include "exodbc/Table.h"
#include "exodbc/ExecutableStatement.h"
#include "exodbc/LogManager.h"
#include "exodbc/ColumnBufferWrapper.h"
#include "exodbc/SqlStructHelper.h"
#ifdef _WIN32
int _tmain(int argc, _TCHAR* argv[])
#else
int main(int argc, char* argv[])
#endif
{
using namespace exodbc;
try
{
// Create an environment with ODBC Version 3.0
EnvironmentPtr pEnv = Environment::Create(OdbcVersion::V_3);
// And connect to a database using the environment.
DatabasePtr pDb = Database::Create(pEnv);
// if argv[1] is given, assume its a connection string, else use some built-in default cs:
// std::string connectionString = "Provider=MSDASQL;Driver={MySQL ODBC 5.3 UNICODE Driver};Server=192.168.56.20;Database=exodbc;User=ex;Password=extest;Option=3;"
// std::string connectionString = ""Driver={IBM DB2 ODBC DRIVER};Database=EXODBC;Hostname=192.168.56.20;Port=50000;Protocol=TCPIP;Uid=db2ex;Pwd=extest;CurrentSchema=EXODBC"
std::string connectionString = u8"Driver={SQL Server Native Client 11.0};Server=192.168.56.20\\EXODBC,1433;Database=exodbc;Uid=ex;Pwd=extest;";
if (argc >= 2)
{
#ifdef _WIN32
connectionString = utf16ToUtf8(argv[1]);
#else
connectionString = argv[1];
#endif
}
// Note: WRITE_STDOUT_ENDL will write to std::wcout on _WIN32, and convert utf-8 to utf-16
// else directly to std::cout
WRITE_STDOUT_ENDL(boost::str(boost::format(u8"Connecting to: %s") % connectionString));
// Open connection to database:
pDb->Open(connectionString);
SqlInfoProperties props = pDb->GetProperties();
WRITE_STDOUT_ENDL(boost::str(boost::format(u8"Connected to: %s, using %s") % props.GetDbmsName() % props.GetDriverName()));
// Create a test table in the database with a few entries
// try to drop first if it already exists, but ignore failing to drop
// auto commit is off by default, so commit changes manually
try
{
pDb->ExecSql(u8"DROP TABLE T1");
pDb->CommitTrans();
}
catch (const SqlResultException& ex)
{
std::stringstream ss;
ss << u8"Warning: Drop failed: " << ex.ToString();
WRITE_STDOUT_ENDL(ss.str());
}
if (pDb->GetDbms() == DatabaseProduct::DB2)
{
pDb->ExecSql(u8"CREATE TABLE T1 ( id INTEGER NOT NULL PRIMARY KEY, name varchar(16), lastUpdate timestamp)");
}
else
{
pDb->ExecSql(u8"CREATE TABLE T1 ( id INTEGER NOT NULL PRIMARY KEY, name varchar(16), lastUpdate datetime)");
}
pDb->ExecSql(u8"INSERT INTO T1 (id, name, lastUpdate) VALUES(101, 'Cat', '1993-10-24 21:12:04')");
pDb->ExecSql(u8"INSERT INTO T1 (id, name, lastUpdate) VALUES(102, 'Dog', '2011-08-01 04:02:06')");
pDb->CommitTrans();
// Create the Table object and Open() it.
// The Database will be queried about a table named 't1' of any type (view, table, ..) during open.
Table t(pDb, TableAccessFlag::AF_READ_WRITE, u8"T1");
t.Open();
// As we did not specify any ColumnBuffer manually before calling Open()
// the Database has been queried about the columns of table 't1'.
// For every column found its SQL Type (like SQL_VARCHAR, SQL_INTEGER) is examined
// to create a ColumnBufferPtr with a matching SQL C Type (like SQL_C_CHAR, SQL_C_SLONG).
// (The mapping of SQL Types to SQL C Types is defined in a Sql2BufferTypeMap and can
// be changed dynamically.)
// For an INTEGER column most probably a LongColumnBuffer has been created,
// while for a datetime column a TypeTimestampColumnBuffer has been created:
auto pIdCol = t.GetColumnBufferPtr<LongColumnBufferPtr>(0);
auto pLastUpdateCol = t.GetColumnBufferPtr<TypeTimestampColumnBufferPtr>(2);
// Depending on the database, driver and operating system, a string column
// might get reported as SQL_VARCHAR or SQL_WVARCHAR. We could either query
// the ColumnBufferPtrVariant about its concrete type, or just ignore
// the concrete type and work with a StringColumnWrapper.
// The StringColumnWrapper will convert from/to string/wstring as needed.
StringColumnWrapper nameColumnWrapper(t.GetColumnBufferPtrVariant(1));
// Lets iterate all rows in the database,
// and print them line by line:
WRITE_STDOUT_ENDL(u8"Printing content of Table 'T1':");
WRITE_STDOUT_ENDL(u8"");
t.Select();
std::string s = boost::str(boost::format(u8"%4s %16s %s")
% u8"id" % u8"name" % u8"lastUpdate"
);
WRITE_STDOUT_ENDL(s);
WRITE_STDOUT_ENDL(u8"==============================");
while (t.SelectNext())
{
s = boost::str(boost::format(u8"%4d %16s %s")
% pIdCol->GetValue()
% nameColumnWrapper.GetValue<std::string>()
% SqlStructHelper::TimestampToSqlString(pLastUpdateCol->GetValue())
);
WRITE_STDOUT_ENDL(s);
}
WRITE_STDOUT_ENDL(u8"");
// Update the row with a primary key id value of 101:
pIdCol->SetValue(101);
pLastUpdateCol->SetValue(SqlStructHelper::InitUtcTimestamp());
nameColumnWrapper.SetValue(u8"Updated");
t.UpdateByPkValues();
// Insert a single entry
pIdCol->SetValue(200);
pLastUpdateCol->SetValue(SqlStructHelper::InitUtcTimestamp());
nameColumnWrapper.SetValue(u8"Inserted");
t.Insert();
// Insert multiple rows using a prepared statement:
ExecutableStatement stmt(pDb);
std::string sql = u8"INSERT INTO T1 (id, name, lastUpdate) VALUES(?, ?, ?)";
stmt.Prepare(sql);
// Bind our columns as parameters:
ColumnBufferPtrVariant pNameCol = t.GetColumnBufferPtrVariant(1);
stmt.BindParameter(pIdCol, 1);
stmt.BindParameter(pNameCol, 2);
stmt.BindParameter(pLastUpdateCol, 3);
// Execute the prepared statement multiple times:
for (int i = 300; i < 310; ++i)
{
pIdCol->SetValue(i);
std::string name = boost::str(boost::format(u8"Insert: %d") % i);
nameColumnWrapper.SetValue(name);
pLastUpdateCol->SetValue(SqlStructHelper::InitUtcTimestamp());
stmt.ExecutePrepared();
}
// Commit everything
pDb->CommitTrans();
// And print the table again:
WRITE_STDOUT_ENDL(u8"Printing content of Table 'T1':");
WRITE_STDOUT_ENDL(u8"");
t.Select();
s = boost::str(boost::format(u8"%4s %16s %s")
% u8"id" % u8"name" % u8"lastUpdate"
);
WRITE_STDOUT_ENDL(s);
WRITE_STDOUT_ENDL(u8"==============================");
while (t.SelectNext())
{
SQL_TIMESTAMP_STRUCT ts = pLastUpdateCol->GetValue();
s = boost::str(boost::format(u8"%4d %16s %s")
% pIdCol->GetValue()
% nameColumnWrapper.GetValue<std::string>()
% SqlStructHelper::TimestampToSqlString(pLastUpdateCol->GetValue())
);
WRITE_STDOUT_ENDL(s);
}
WRITE_STDOUT_ENDL(u8"");
}
catch (const Exception& ex)
{
WRITE_STDERR_ENDL(ex.ToString());
}
catch (const std::exception& ex)
{
WRITE_STDERR_ENDL(ex.what());
}
return 0;
}
exOdbc was written by Elias Gerber.
Contact: eg (AT) elisium.ch