We hit this exact Data Table case-sensitivity issue when integrating GC with an AWS Lambda function.
Our Lambda writes the customer tier (GOLD, SILVER, BRONZE) into a Data Action response. The Architect flow looks up that value in a Data Table. The Lambda returned Gold (title case), but the Data Table key was GOLD (uppercase). The lookup returned null. We fixed it by normalizing to uppercase in the Lambda before responding.
# Lambda normalization
return {'tier': customer_tier.upper()}
This bites me on every single client deployment.
Callers enter their account number via DTMF, and the flow looks it up in a Data Table. But the DTMF input sometimes captures a leading space or trailing newline character that is invisible in the Architect debugger. The lookup fails, the caller gets the generic treatment, and the containment rate tanks. I always wrap the input in Trim(ToUpper(input)) before every Data Table lookup.
We had this same headache when migrating Zendesk macros to GC Data Tables.
Zendesk stores tag values in lowercase by convention. When we exported the tags and imported them as Data Table keys, all the keys were lowercase. But our Architect flow was comparing against the original mixed-case values from the CRM. Three days of debugging before we realized it was just a casing mismatch.
If you bulk-load Data Tables via the API, double-check your CSV encoding.
I’ve seen ETL pipelines that export Data Table rows from a source system with UTF-8 BOM encoding. The invisible BOM character (U+FEFF) gets prepended to the first key in the CSV. The Data Table ingests it silently, but the key now starts with an invisible character, making every lookup against that first row fail.
This is a documented behavior that catches many implementers off guard.
- Data Table lookups are case-sensitive. The key
GOLD ≠ Gold ≠ gold.
- Architect string comparisons are also case-sensitive by default.
- Use the
ToUpper() or ToLower() expression function in Architect to normalize before lookup.
Please reference the Resource Center article: ‘Working with Data Tables in Architect flows’.