Neuroshok Database Interface

Master	Unix
Dev	Unix

ndb is a generic interface to connect to any databases. Querys are written in pure C++ and can be executed on different database engines just by changing one parameter. Database model is defined and accessible compile time.

Features

• Header only interface
• Full C++ and generic querys
• Support any type of databases
• Easy to add new engines
• Compile-time generation for SQL expressions
• Customizable types conversion
• Field access by C++ structure
• Database generation with macros

Current support

• Expressions based on SQL and BSON
• Sqlite, MySQL and MongoDB.

How to use

Choose the branch to clone (master or dev)

ads@NK:/home/ads$ git clone --recurse-submodules -b [branch] https://github.com/ads00/ndb.git

Build (optional)

Create build directory

ads@NK:/home/ads/ndb$ mkdir build && cd build

Bundled engines

Choose engines with option -DNDB_ENGINE_[name]=ON (ex: -DNDB_ENGINE_SQLITE=ON)

Options

-DNDB_BUILD_TEST=ON -DNDB_BUILD_EXAMPLE=ON

ads@NK:/home/ads/ndb/build$ cmake -DNDB_ENGINE_SQLITE=ON [options] -DCMAKE_BUILD_TYPE=Release ..
ads@NK:/home/ads/ndb/build$ make
(optional) ads@NK:/home/ads/ndb/build$ make test

Integration

CMake

add_subdirectory(${PATH_TO_NDB}/ndb ${THIRD_PARTY_ROOT}/ndb/cmake-build)
target_link_libraries(my_target lib_ndb)

Remember, it's importent to enable options like engines before adding the subdirectory. Indeed, you need to include libraries of the engine you use.

Manual

Add paths to ndb headers and your customs engine builds

Overview

Database

ndb_table(movie,
          ndb_field_id,
          ndb_field(name, std::string, ndb::size<255>)
)
ndb_table(music,
          ndb_field_id,
          ndb_field(image, std::string, ndb::size<255>)
)
ndb_model(collection, movie, music)

ndb_project(my_project,
            ndb_database(libray, collection, ndb::sqlite),
            ndb_database(mongo_library, collection, ndb::mongo)
)

Querys

ndb::query<dbs::zeta>() << ( movie.id, movie.image ); // get
ndb::query<dbs::zeta>() << ( movie.id == a && movie.name == b ); // get by condition
ndb::query<dbs::zeta>() >> (( movie.name = "updated") << ( movie.id == 3 )); // update by condition
ndb::query<dbs::zeta>() + ( movie.id = 3, movie.name = "test" ); // add
ndb::query<dbs::zeta>() - ( movie.id == 3 ); // del

Example

An example with a libray database using a collection model and a movie table

#include "my_database.h"

// aliases
namespace dbs
{
    using libray = ndb::databases::my_project::libray_;
    using mongo_library = ndb::databases::my_project::mongo_library_;
}

int main()
{
    // alias
    const auto& movie = ndb::models::collection.movie;
    
    // initialize sqlite and mongo
    ndb::initializer<ndb::sqlite, ndb::mongo> init;
    
    // connect to database library (sqlite is used)
    ndb::connect<dbs::library>();
    
    // connect to database mongo_library (mongo is used)
    ndb::connect<dbs::mongo_library>();  

    // add a movie with specified movie.name and movie.duration (table is deduced compile time)
    ndb::query<dbs::library>() + (movie.name = "Interstellar", movie.duration = 2.49_h) );
    
    // get all fields from movie where duration is less than 2H30
    for (auto& line : ndb::query<dbs::library>() << (movie.duration <= 3.30_h))
    {
        std::cout << "movie.name : " << line[movie.name] << std::endl;
        std::cout << "movie.duration : " << line[movie.duration] << std::endl; 
    }

    return 0;
}

Documentation

Example

Contribution

ads00

Jonathan Poelen (@jonathanpoelen)