detule / cyanodbc

Cython wrapper for Nanodbc

cyanodbc

A Cython Wrapper for Nanodbc which dubs itself as a "A small C++ wrapper for the native C ODBC API".

Build Status

Branch	Linux/OSX	Windows	Coverage
master

Building

To build Cyanodbc, you'll need CMake, Ninja, Cython, and Python. Additionally to run tests you will need pytest. The good news is all of these are available on PyPi, so they are just a pip install away.

pip install requirements.txt

On all platforms the first step is cloning the repo and seting up the build directory

git clone --recurse-submodules https://github.com/cyanodbc/cyanodbc.git
cd cyanodbc
mkdir build
cd build

From there, the next step is to build the shared library, using the cmake build system:

cmake -G Ninja -DNANODBC_ODBC_VERSION=SQL_OV_ODBC3 -DCMAKE_BUILD_TYPE=Release -DCYANODBC_TARGET_PYTHON=3.5 ..
cmake --build .

Cmake notes:

• Make sure the variable DCYANODBC_TARGET_PYTHON is set to the corresponding python version you are using - this works around certain quirks of CMake python identification.
• Nanodbc is built and linked automatically when you build cyanodbc. We expose the following cmake compile-time options:

CMake Option	Possible Values	Details
NANODBC_ENABLE_BOOST	OFF or ON	Use Boost for Unicode string convertions (requires [Boost.Locale][boost-locale]). Workaround to issue #24.
NANODBC_ODBC_VERSION	SQL_OV_ODBC3[...]	Forces ODBC version to use. Default is SQL_OV_ODBC3_80 if available, otherwise SQL_OV_ODBC3.

Please check the Nanodbc Project for more information on these as well as other options that can be made available, if there is a need.

Finally, once the shared library is built, you can install the package using pip

pip install -e src/cython

Testing

Pytest is used for running the tests.

simply run

cd cyanodbc
pytest tests