A partial unique index allows to constrain only a subset of the records in a table. This post shows an example how to implement a partial unique constraint with SQLAlchemy.

Upgrading old NixOS machines can lead to a situation where the rebuild fails due to an unknown hash algorithm. This post shows a workaround to get started again.

Setting custom meta data with Pelican can be implemented easily. This post shows a small example how I solved it.

Small example how to print out a SQLAlchemy query with parameters being expanded.

When displaying items based on a database search, one sometimes wants to display a specific set of items at the top or bottom of the result set. Applying the custom order inside of the database query allows to still use LIMIT and OFFSET in a natural way.

A hint on how to store the Heroku Connect mapping configuration in a repository, so that diffs are readable and a useful information of the system changes over time.

Two approaches to quickly compare two similar tables and get an idea of how many records differ based on an example with geo data.

Thoughts on how an integration of a CLI (command line interface) toolkit for the Pyramid web framework might look like and could be integrated into the configuration mechanisms of Pyramid itself.

Working with Heroku Connect means that at some point some records will end up in a failed state. This post shows a few options how to trigger the sync manually again.

Channels can play a tricky role when using the Nix package manager to manage development environments. Especially when one want to make sure that a derivation builds against multiple channels. This post shows how to quickly build a derivation against a different channel based on setting a URL into the include path for Nix.