何时使用Securefile LOB重复数据删除和压缩？

Question

I have a table with BLOB column and SDO_GEOMETRY column (Spatial data). The BLOBs are stored as securefiles (Oracle 11g) I may be having quite a few repetitive columns of BLOB data. I am contemplating about adding compression or deduplication or both to save storage. I am wondering what is the difference between compression vs deduplication (compress is also going to save space by not storing duplicate data ?) When is it better to use them?

Answer 1

I don't think I can explain the difference better than this Oracle Technology Network artice , which among explaining what compression and deduplication are for, says:

Compression is not the same as deduplication. Compression happens inside a LOB column, inside a row—each LOB column is compressed independently. In deduplication, all the rows are examined and duplicate values in the columns are removed and replaced with pointers. If you have two very different rows, deduplication will not reduce the size; but compression may optimize the space inside the LOB value. You can compress as well as deduplicate the table.

So if you have the same BLOB in multiple rows deduplication will save space. Compression is more likely to be useful for individual CLOB values, but whether it's useful for your BLOBs depends on what is in them - if they are already in a compressed format (eg JPEG, as the article mentions) then compressing the BLOB won't save any space, and Oracle won't waste much time trying.