Hi Matt, In addition to Martins comments, aliases are a good way to link different names together. They can be used for schema evolution, but also for projecting a table schema onto an avro record.
Note: aliases can be anything (they are intentionally not verified). The only constraint is that they must be unambiguous / unique within their scope. Kind regards, Oscar -- Oscar Westra van Holthe - Kind <os...@westravanholthe.nl> Op za 12 okt. 2024 10:56 schreef Martin Grigorov <mgrigo...@apache.org>: > Hi Matt, > > Which SDK do you use ? > > Most of the SDKs support custom attributes so you should be able to store > the invalid names in them and later recover the initial values! > > Some of the SDKs allow the use of custom name validators. This way you can > relax the checks. > For the Rust SDK see > https://github.com/apache/avro-rs/blob/main/avro/README.md#custom-names-validators > > IIRC the Java SDK also has hooks for this but I can't point you to the > exact APIs at the moment. > > Martin > > On Thu, Oct 10, 2024 at 8:14 PM Matt Burgess <mattyb...@gmail.com> wrote: > >> Hi all, >> >> I'm new to this list so I apologize if this has been discussed before, >> but I couldn't find anything in the archives about it. >> >> I'm on the PMC for Apache NiFi and I'm facing an issue with a source and >> a target database that allow digits as the first character of a table / >> column name. I'd like to be able to keep the data in Avro throughout the >> NiFi flow, but the names are not valid for Avro. Is there any workaround >> (perhaps using aliases or custom attributes/metadata?) for having the >> invalid name somewhere in the schema that my code could use (while renaming >> the fields to Avro-valid names in the middle of the flow) to "restore" the >> original name when inserting the data into the target database? >> >> Thank you in advance, >> Matt >> >
