Again for tap-postgres target-postgres, Can I configure FULL_TABLE refresh only if table doesn't exists in target and if it exists do not drop just skip. If it's possible, will the common tables still add an additional load on source? Will it still be queried even though not written to the target?

Can I configure FULL_TABLE refresh only if table doesn't exists in target and if it exists do not drop just skip.

Technically yes, but you'd have to write the logic. You can use https://docs.meltano.com/reference/command-line-interface#select Selects to select which tables you want to sync. Taps/targets don't know about each other by design (except for state) I think a why would at least help with formulating a feature request, backfill probably?

If it's possible, will the common tables still add an additional load on source?

I don't know what this means

Will it still be queried even though not written to the target?

I'm not sure what this means, but yes if data is pulled from the tap the target updates the data. I think it would be helpful for you to read up on https://hub.meltano.com/singer/spec , and play with

meltano invoke tap-postgres > output

cat output | meltano invoke target-postgres

and look at the output file in between messing with the combinations you're asking about

Taps/targets don't know about each other by design (except for state)

this is enough to clarify, thanks.