CustomDimensions that contain letters not in the English alphabet

We send CustomDimensions in scope visit, that contain letters that are not part of the English alphabet, such as “ø” or “å”. This seems to cause Matomo to split the strings before the “unknown” letter.

An example is a CustomDimension named “Trøndelag”, which in Matomo is shown as only “Tr” in some cases.

Anyone familiar with this problem? Or know what we can do to avoid that the strings are cut before the letter “ø”?

How is configured your database (in term of charset)?

My database has not been configured other than connecting my Matomo pod to it in our local onprem Kubernetes environment. I am using the MySQL server 5.7 image (https://hub.docker.com/layers/mysql/mysql-server/5.7/images/sha256-5278734456ac1771457c620684c1e80612743c0be06b681ed6649d72a80102fa?context=explore) and done nothing more than to deploy it as a statefulset in Kubernetes.

The specific issue is the Norwegian charset, specifically “å” and “ø”. Though the issue doesn’t seem to be consistent as both “å” and “ø” are captured by Matomo in some cases, but in other cases it doen’t seem to be caught. Example is a word “Sør” that is either logged at a Matomo CustomDimension “S” or “Sør”. Both values come through.

Hi @betinapbye
Could you post some screenshots where it works and where it doesn’t?