If there is any compression going on you don't really have a choice. Long ago I have actually done code that updated files (one field per record, no size changes) in a .zip file but the creation specifically forced the files in question to be stored with no compression.
(Legacy evolution--all the files of a job were packed into a single .zip. Then a need came along to update some of them. Nothing else would even notice that the files hadn't been compressed, but the one task could very quickly write the changes without replacing the file. Update the relevant byte, recalc the CRC from the in-memory copy, write it to both locations.)
(Legacy evolution--all the files of a job were packed into a single .zip. Then a need came along to update some of them. Nothing else would even notice that the files hadn't been compressed, but the one task could very quickly write the changes without replacing the file. Update the relevant byte, recalc the CRC from the in-memory copy, write it to both locations.)