Hi all, the page.json file being shipped with Spryker defines a dynamic template for string sorting fields setting the type to “keyword”. Hence sorting of umlauts (and other special chars) is odd and does not behave like naturally expected in a german locale. Changing the type from “keyword” to “icu_collation_keyword” solves this problem for us. However, the Elasticsearch instance being shipped with Docker SDK does not have the analysis-icu plugin installed and that is where the problem kicks in: Is there a native way of installing a plugin for Elasticsearch in Docker SDK?

Hello Dennis, Your problem is similar to this one: https://sprykercommunity.slack.com/archives/CMYFJCG2E/p1597243552296300 The recommendation are the same.