Im trying to remove all columns that start with some prefix from all rows, its taking a really long and Im trying to optimize it by not fetching the values since they can be heavy.

e.g., if I have a row with key1 and column family test that has columns a1, a2 and d, I want to remove all columns that start with a so a1 and a2.

I tried looking into delete by prefix / filter but its not supported, and I saw KeyOnlyFilter and FirstKeyOnlyFilter but they only give me the row key, is what Im trying to do possible?

Im using the Java API, but if there is something thats only available from the hbase shell its also an option.