Closed gschmutz closed 1 year ago
Ok, typical Monday error ;-) of course my problem with the id
can easily be solved using the pk.mode=record_key
. So I'm closing this issue. Such an option to keep fields might be interesting for other cases, but it would also make your SMT more complex ....
Hi
Have an other problem where I would like to hear your opinion. I have the following structure, where I have created the extra "ID" field using a
com.github.jcustenborder.kafka.connect.transform.common.ExtractNestedField$Value
before thecom.github.cedelsb.kafka.connect.smt.Record2JsonStringConverter$Value
SMT:I wanted to use the extra
id
field as the primary key in the JDBC sink connector. But of course the Record2JsonStringConverter SMT is taking the whole struct and transforms it into a single field.Now I'm wondering if you would support adding an option to "blacklist" some fields not be included in the transformation and by that left as is. It could be named
...keep.fields
. Or do you see any other way to do it?Having this
id
field besides the "jsonstring" would allow me to use theupsert
mode together with thepk.fields
option of theJdbcSinkConnector
.Again I could try to implement it and send you a PR.