batyr: On-demand synchronization of vector data to a PostGIS database

One common situation when dealing with geographic data is repeatedly exporting and importing this data to and from a PostGIS-enabled database. While the export is very well covered by products like Mapserver and GeoServer, importing is a bit more tricky. Common solutions consist mostly of custom scripts wrapping commands like shp2pgsql or ogr2ogr. These solutions often fail or at least need some tricky hacks if single rows of data should be updated instead of deleting and restoring the complete table content. It is also hard to account for slow or interrupted transactions and still make sure that the data stays synchronized as a whole. Using these import scripts requires either command line access or some custom code to hook them up to a job queue or even web interface to make them usable from within other applications.

Flaws like those were the reason for us to create batyr as a reusable solution for similar demands in the future.

status overview

batyr is a single server application providing the following:

  • "Intelligent" writing of data. A synchronization does not consist of a complete truncate and restore of a table anymore. Only features which have any differences to the ones provided by the external datasource are actually updated. New features are only created if they are not already in the database and features get (optionally) removed from the database if they are not part of the datasource any more. All this uses the primary key of the table to identify matching features from the datasource.
  • An integrated web-interface to get an overview on the current state of the server and to optionally start syncronizations manually.
  • A well-documented HTTP-API to easily integrate the batyr into other applications and allow flexible triggering of synchroniszations. Furthermore the HTTP-API provides methods to integrate batyr in existing monitoring systems like Nagios.
  • On-the-fly transformation of geometries to the spatial reference system of the database table. The required reference system is looked up in the PostGIS geometry_columns view/table and the transformation itself is performed by PostGIS.
  • Internally batyr uses the OGR-library to access datasources. So batyr covers all vector formats supported by OGR and connecting to - for example - a WFS is possible. Additionally this allows using OGR Virtual Formats for extended configuration options.
  • Synchronization jobs are internally queued and are handled in parallel using a configurable number of database connections. This takes care of a responsive HTTP-API as well as optimal usage of resources.

With these features it is possible to quickly integrate external geodata into your PostGIS database - without having to spend time creating custom code.

job queue

The batyr-project has been sponsored by the trafimage project of the swiss federal railways (SBB). The software is open-source and available on github.

More on this topic
8 min reading time › | Blog

Adding type hints to existing code in Python

The Python interpreter handles types in a dynamic and flexible way without constraints on what type of object a variable is assigned to. Since Python 3.5 programmers have the option to add type annotations to their code to check whether variable types are valid. In this blog entry we show you how it's done.

read more
2 min reading time › | Blog

Next.js Routing with Nginx

Automatically generate a Nginx routing configuration for static Next.js projects.

read more
2 min reading time › | Blog

From backend to frontend: we are looking for reinforcement

We have two vacancies for Fullstack or Backend Developers for our offices in Freiburg and Olten. If you want to bring your passion to a strong team, then you should apply.

read more
3 min reading time › | Blog

Migrating from enzyme to testing-library/react

We have rewritten our frontend unit tests from using enzyme to testing-library/react. This article provides a quick overview of the updates.

read more
2 min reading time › | Blog

mapset 2.0 with MUI

mapset 2.0 has been widely refactored using the open-source react component library MUI (Material-UI).

read more
3 min reading time › | Blog

Set up Django to only allow CORS requests in DEBUG mode

This post is about how to set up a Django project to only allow CORS requests in DEBUG mode, even if they require a login to the backend. In our case, this has been useful to test frontend customizations on the internal dev environment without having to start the backend locally.

read more


geOps AG
Solothurnerstrasse 235
CH-4600 Olten

fon: +41 61 588 05 05
geOps GmbH
Bismarckallee 10
D-79098 Freiburg

fon: +49 761 458 925 0