We have our custom columnar database. So I wrote a custom bulk loading step to load transformation data onto our database. Now we are looking to do clustering and I was wondering if anyone could tell me how to achieve this.I.e say for postgres we can specify a set of connections as clusters and specify it in the Table output step. I am looking at something similar to that . I was wondering what components would have to be overridden to achieve this ? Any help would be appreciated. Thanks in Advance.