Postgres C++ interfacing library : pgi

pgi is a header only library built on top of libpqxx and cpp-yaml. It is essentially a wrapper around libpqxx that allows configuration from yaml file and higher level methods interfacing C++ with postgres.

Example

    DatabaseWorker dbw("database_config.yaml");

    std::map<std::string, double> double_cols;
    double_cols["test_double1"] = 1.1;
    
    std::map<std::string, std::string> string_cols;
    string_cols["test_string"] = "Hello world!";
    
    std::map<std::string, time_point_t> time_cols;
    time_cols["time"] = tp;
    
    dbw.insert_from_maps("public.test_table", double_cols, string_cols, time_cols);
    dbw.print("public.test_table");

Requirements

libpqxx installed
cpp-yaml installed

Installation

Installation is done using cmake. At the root of the source directory, run

mkdir build && cd build && cmake .. && make install

How to use

A full standalone example can be found in the example/ directory. Follow those steps to reproduce :

In CMakeLists.txt of your project, import the pgi library like so :

cmake_minimum_required(VERSION 3.10)
find_package(pgi CONFIG REQUIRED)
add_executable(main main.cpp)
target_link_libraries(main pgi)

Create a config file with yaml format. An example is given in config/database_config.yaml. See config file fields for details.
Include <pgi.hpp>
Create a DatabaseWorker instance, by providing a path to the configuration file.
Use DatabaseWorker methods. For example, insert columns from multiple std::map and print the content of the table :

    DatabaseWorker dbw(
        pgi_test::pgi_config_dir + "database_config.yaml", pgi_test::pgi_config_dir + "database_config_out.yaml");

    // Most general method to insert a row
    std::map<std::string, double> double_cols;
    double_cols["test_double1"] = 1.1;
    double_cols["test_double2"] = 2.2;
    double_cols["test_double3"] = 3.3;
    std::map<std::string, std::string> string_cols;
    string_cols["test_string"] = "Hello world!";
    std::map<std::string, time_point_t> time_cols;
    time_cols["time"] = tp;
    time_cols["time2"] = tp - 50s;
    dbw.insert_from_maps("public.test_table3", double_cols, string_cols, time_cols);
    dbw.insert_from_maps("public.test_table3", double_cols, string_cols, time_cols);
    dbw.insert_from_maps("public.test_table3", double_cols, string_cols, time_cols);
    dbw.print("public.test_table3");

Try it with docker-compose

A minimal working environment (for the sake of example + continuous deployment) can be found in ci/. It sets up a minimal postgres database in one container, builds minimal example using pgi in another container and runs it.

./run_ci.sh

Will end with output :

test_double1 |test_double2 |test_double3 |time                    |test_string  |time2                   |
1.1          |2.2          |3.3          |2021-12-01 11:32:38.112 |Hello world! |2021-12-01 11:31:48.112 |
1.1          |2.2          |3.3          |2021-12-01 11:32:38.112 |Hello world! |2021-12-01 11:31:48.112 |
1.1          |2.2          |3.3          |2021-12-01 11:32:38.112 |Hello world! |2021-12-01 11:31:48.112 |

Configuration file fields

connection : have to contain all the fields to establish database connection. Key/Values will be parsed to the connection string. See postgres documentation for the connection string here.
[tables] : list tables to explore at the construction of DatabaseWorker
[field_length_mapping] : configure the length of the printed column according to postgres type