We tried to replicate the scenario you described, and we’re currently not experiencing any issues with querying metadata values that include percent-encoded escape characters, such as the hash sign (“#”).
Upon thorough testing on our end, we have confirmed that using percent-encoding, with the hash sign replaced by %23 in the URL, works correctly with a correct response.
Thank you for a quick reply.
Sorry I didn’t mention this at first, but for me it didn’t work when hash sign is the first or the last character in a metadata.value string. Could you also check this case?
Finally I understand my mistake.
I was using Postman to create those request and in my case the problem was with a “+” sign that Postman didn’t convert it automatically to %2B.
So my advice would be just to convert all special characters to “Percent encoding”.