Skip to main content

Hi Airtable team,

I've encountered a Unicode-related issue with the “contains” filter in Airtable that leads to inconsistent and unexpected behavior when searching for certain characters – specifically, characters like “ü” stored as composed Unicode sequences.

🧪 Reproduction Scenario:

I'm importing records from an external API into Airtable. These records include names with the word “Vorzüge”, but the “ü” character is stored as a composed form:
u (U+0075) + combining diaeresis ¨ (U+0308) – rather than as a single precomposed character ü (U+00FC).

Example records:

  • 📂 Vorzüge record A (with composed “ü”: u + ¨)

  • 📂 Vorzüge record B (with normal/precomposed “ü”)

🧩 Problem:

  • When using the "contains" filter on the field and typing "Vorzüge", only the record with the precomposed “ü” (U+00FC) is matched.

  • The record with the composed form (U+0075 + U+0308) is not found, even though the text appears identical.

  • However, the global Airtable search (e.g., Cmd/Ctrl + K or in-table search) does return both records, suggesting that search uses Unicode normalization – while filters do not.

✅ Expected Behavior:

The “contains” filter should match Unicode-equivalent characters, even when they are technically stored using different code points (e.g., NFC vs. NFD normalization forms). If two characters look and behave the same, users reasonably expect them to match in filters.

Hey ​@Marks_GmbH_Nord,

This is interesting. You might want to reach out via emsupport@airtable.com or fill out this form (technically for Product Ideas) but I believe that it is still a good place to bring this up.

 Mike, Consultant @ Automatic Nation

Hm, was curious about this and attempted to replicate via the ‘Example records’ data and filtering by contains with the  "Vorzüge" bit, but that worked fine for me.  Not saying there isn’t an issue or anything, just wanted to provide a data point!

 

Hm, was curious about this and attempted to replicate via the ‘Example records’ data and filtering by contains with the  "Vorzüge" bit, but that worked fine for me.  Not saying there isn’t an issue or anything, just wanted to provide a data point!

 

Thank you very much for testing.

  • Did you also check whether the "Ü" in record A is actually a composed "Ü"?

  • Did you copy the text of both records from the post? There was no composed “ü”

  • Please copy this word from this post into your data: Vorzüge – it contains a composed "Ü".

For now, we've added a formula check field to validate our data.

 

This Unicode normalization issue definitely impacts accurate search results, especially in systems handling diverse text inputs. For anyone needing precise and up-to-date info, the Arrest lookup VA offers a reliable resource for real-time arrest data and can be a good example of how consistent search functionality is crucial for user experience. Checking how different platforms handle similar challenges can provide valuable insights for improving filter behavior.

Yes, absolutely. That’s why it’s extremely important for Airtable to make improvements in this area.

Reply


Terms & Conditions