2 ms·
The image near the top of the readme implies that compressing the whole file means not having random access, which is indeed true, but it also implies that ther
by cryptonector 4y ago
The image near the top of the readme implies that compressing the whole file means not having random access, which is indeed true, but it also implies that there's nothing between whole file compression and row/column level compression, and that is not true. You could compress pages, kinda like ZFS compresses blocks, and still maintain random access, and get pretty close to whole-file compression.
- dataangel 4y agoThat requires changing your filesystem to a compressing one though, and you lose the compression when you copy the file to another filesystem. It does have the advantage that the page cache could have the uncompressed copy though.
- cryptonector 4y agoNo, the database itself can compress in the same way as ZFS. Now, don't layer compression on compression, otherwise it should work.