Hi All, I am creating a connector for a system tha...
# contributing-to-airbyte
w
Hi All, I am creating a connector for a system that can also list the deleted rows from a certain point in time. Is there a way to mark rows as deleted in an incremental stream? Would it work if I add ab_cdc_deleted_at and ab_cdc_updated_at columns to my output?
Hi @Wisse Jelgersma, do you want deleted rows to not be part of the normalized output or do you simply want to flag them as deleted?
We would prefer it be removed in the final table. This is what happens with CDC sources as well, right?
Yes, I think so. @Chris (deprecated profile) I believe this is an exciting normalization question: will basic normalization for CDC work if the source connector used is not a JDBC-based source? Is a source considered as CDC only according to its schema?
Hi, I have a same question that I want to remove rows that are 30 days old. What would be preferable way, when using airbyte with postgres?
u
@Zawar khan will transfer your data from your source to destination, so one way is to use a view that only has 30 days and use full refresh (in case the table is not large) other way is to use dbt to post-transform your data.
b
@[DEPRECATED] Augustin Lafanechere I see @Chris (deprecated profile) is on holiday, is there anyone else in the Airbyte team that can answer my question in the meantime? I think you formulated my question perfectly.
c
Would it work if I add ab_cdc_deleted_at and ab_cdc_updated_at columns to my output?
Yes, that should probably work
m
Thanks @Chris (deprecated profile), I will try it out
s
So far it looks like it's working! thanks @Chris (deprecated profile) and @[DEPRECATED] Augustin Lafanechere
Thank you @Wisse Jelgersma for the experimentation 😄