Why does snappy encodes the file ( increases the size ) instead of compressing it ? about snappy HOT 1 OPEN

Any compression algorithm that reduces the length of some strings will necessarily increase the length of others (intuitively, there are fewer possible codes of length n than strings of length n+1). Typically random-looking data will result in at least some expansion, since each literal will have some framing and Snappy doesn't use entropy coding.

I wouldn't expect more than a couple percent overhead even from a random file, but some expansion is possible.

from snappy.