Skip to content
SQL Programming Assistant for PostgreSQL
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

The SQL Programming Assistant for PostgreSQL

The SQL Programming Assistant for PostgreSQL (hereinafter referred to as Pgspa) - is a command line program and the PostgreSQL extension to simplify the development of the PostgreSQL databases.

ATTENTION, this software is "alpha" quality, use it at your own risk!

Pgspa allows to execute a set of SQL queries from files of arbitrary directory hierarchy in one transaction. The queries of the set are executed iteratively. Queries ending with the ignorable errors are not executed in the next iteration, while queries ending with the non-fatal errors are executed in the next iteration. The execution stops when there are no successfully executed queries, or there are at least one query ended with a fatal error during the iteration. Finally, if not all the queries of the set considered done, the entire transaction is rolled back and Pgspa reports about the problems.

Table 1. Ignorable errors (queries ended with these errors are considered done)

SQLSTATE code Condition name
42723 duplicate_function
42P06 duplicate_schema
42P07 duplicate_table
42710 duplicate_object

Table 2. Non-fatal errors (queries ended with these errors will be re-executed in the next iteration)

SQLSTATE code Condition name
2BP01 dependent_objects_still_exist
3F000 invalid_schema_name
42883 undefined_function
42P01 undefined_table
42704 undefined_object

All other errors are treated as fatal and leads to failure of the entire transaction.

Pgspa also ships with the server-side PostgreSQL extension dmitigr_spa which provides the convenient set of functions for database developers including functions to drop the interdependent database objects in the non-cascade mode.


To start working with Pgspa the directory with the SQL source files must be marked as the "project directory" by using pgspa init command:

$ cd /path/to/the/database/project
$ pgspa init

This command will create the directory called .pgspa.

The project directory and its subdirectories can contains the following files:

  • the SQL sources - files with ".sql" extension;
  • the shortcuts - files without extension;
  • .pgspa_config - per-directory configuration file.

The SQL source files, directories and shortcuts are so called references. The reference name cannot be empty or starts with the dot (".").

The extension for PostgreSQL can be created by using the following SQL query:


SQL source files

Each SQL query (except the very last) must ends with the semicolon. At this moment, SQL queries cannot be parameterized, but there are plans to add this feature in the future.

The SQL source files can be organized in the arbitrary directory hierarchies. They will be executed in lexicographical order of the file names. Each directory can contain the both file foo.sql (so called heading file) and directory foo/. In this case, the command pgspa exec foo will execute the queries of foo.sql at first, and then the queries of foo/.

The concrete SQL file and/or set of SQL files and directories can be specified as the arguments of the exec command, for example:

$ pgspa exec foo bar.sql baz

All queries of the specified bunch will be run in the same transaction. Since Pgspa is using GNU style for reporting error messages, Emacs users can run Pgspa in the Compilation Mode to immediately jump to the erroneous queries.


Shortcuts - are regular files without an extension that contains the names of SQL files, directories and/or shortcuts (i.e. references) relative to the directory in which the shortcut is located. Shortcuts are useful to define a sequence of references to be executed without the need to specifying them in the command line each time. For example, consider the following simple case:


Here obviously, that create.sql contains the DDL queries to create the objects of the schemas, and drop.sql contains the DDL queries to drop the objects of the schemas. To recreate the objects, the following command can be used:

$ pgspa exec schemas/drop schemas/create

The same effect can be achieved by creating the shortcut recreate - a file named as recreate with the following content:


Then this shortcut can be used like that:

$ pgspa exec schemas/recreate

Note, that cyclic shortcuts are not allowed.

But there are still some gotcha: all queries of the whole schemas directory can be accidentally executed as simple as:

$ pgspa exec schemas

It will lead to executing schemas/create.sql first and then schemas/drop.sql. It's absolutely senselessly and may be even dangerous, since all of the objects will be eventually dropped. Such incidents can be prevented by using per-directory configuration (namely, the parameter explicit, described below).

Per-directory configuration

Each subdirectory of the project directory can contain the configuration file named .pgspa_config. The format of this file is very simple:

parameter1 = value_without_spaces
parameter2 = 'value with spaces'
# commented_out_parameter_with_empty_value =

Currently supported parameters are:

  • explicit - is a boolean parameter (possible values are "yes"/"no") which allows to run only SQL queries by the explicitly specified references. For example, if the directory foo is marked with explicit parameter, the only way to run the SQL queries of this directory is to use one of the reference of this directory, like foo/bar or foo/baz.sql.



The table below (may need to use horizontal scrolling for full view) contains variables which can be passed to CMake for customization.

CMake variable Possible values Default on Unix Default on Windows
The type of the build
CMAKE_BUILD_TYPE Debug | Release | RelWithDebInfo | MinSizeRel Debug Debug
Installation directories
CMAKE_INSTALL_PREFIX an absolute path "/usr/local" "%ProgramFiles%\dmitigr_pgspa"
DMITIGR_PGSPA_PG_SHAREDIR an absolute path not set (can be set manually) not set (can be set manually)


Pgspa ships with the PostgreSQL extension called dmitigr_spa which consists of two files:

  • dmitigr_spa--0.1.sql
  • dmitigr_spa.control

It will be placed to the directory specified via the DMITIGR_PGSPA_PG_SHAREDIR variable. (Usually, it should be something like /usr/local/pgsql/share/extension on Linux and C:\Program Files\PostgreSQL\10\share\extension on Microsoft Windows.)

The default build type is "Debug".

Installation on Linux

$ git clone
$ mkdir -p pgspa/build
$ cd pgspa/build
$ cmake -DDMITIGR_PGSPA_PG_SHAREDIR=/usr/local/pgsql/share ..
$ make
$ sudo make install

Installation on Microsoft Windows

Run Developer Command Prompt for Visual Studio and type:

> git clone
> mkdir pgspa\build
> cd pgspa\build
> cmake -G "Visual Studio 15 2017 Win64" -DDMITIGR_PGSPA_PG_SHAREDIR="C:\Program Files\PostgreSQL\10\share" ..
> cmake --build --config Debug .

Next, run the elevated command prompt (i.e. the command prompt with administrator privileges) and type:

> cd pgspa\build
> cmake -DBUILD_TYPE=Debug -P cmake_install.cmake


Pgspa is distributed under zlib license. For conditions of distribution and use, see file LICENSE.txt.


Contributions and any feedback are highly appreciated! Donations are welcome!


Copyright (C) Dmitry Igrishin

You can’t perform that action at this time.