How to perform accent/diacritic-insensitive search on text fields?

General
1

I’m building a multilingual application with content in languages that use diacritics (Czech, French, etc.), and I need to implement search functionality that ignores accents/diacritics.

Current Situation:

  • Users can search blog posts by title and perex (excerpt)

  • I’m using _icontains filter operator for case-insensitive search

  • The search should match “cafe” with “café”, “Přemysl” with “Premysl”, etc.

Questions:

  1. Is there a built-in filter operator in Directus that performs accent-insensitive search (similar to PostgreSQL’s unaccentor MySQL’s collation)?

  2. What’s the recommended approach for this use case?

    • Should I store normalized versions of searchable fields?

    • Can I use custom database functions with Directus filters (i’m using Sqlite as storage)?

    • Is there a better pattern for multilingual search?

My current filter:

filter: {
_or: [
{ “title”: { “_icontains”: searchText } },
{ “perex”: { “_icontains”: searchText } }
]
}

Any guidance would be greatly appreciated!

Thank you

1 Like