[WIP][POC] Add FSST encoding - #3784

Draft

prtkgaur wants to merge 10 commits into

apache:masterfrom

prtkgaur:fsstEncoding

Commits on Sep 8, 2026

Port the FSST symbol table trainer and compressor to Java
sfc-gh-pgaur
committed
141d6fc
Add the symbol table data page body and the table handoff seams
sfc-gh-pgaur
committed
52b0461
Encode and decode pages against a symbol table
sfc-gh-pgaur
committed
d0607bc
Let a column ask for the symbol table encoding
sfc-gh-pgaur
committed
b385eda
Assert with JUnit 5 and AssertJ, as the rest of the module does
sfc-gh-pgaur
committed
f9a7350
Decode symbol table pages written by another implementation
sfc-gh-pgaur
committed
e64fad2
Carry the symbol table from writer to reader through the column machinery
sfc-gh-pgaur
committed
2667916
Add an end-to-end test for the symbol table transport
sfc-gh-pgaur
committed
6494e81
Check the offset section in the encode direction against the fixtures
sfc-gh-pgaur
committed
e41501e
Benchmark FSST against the encodings a text column falls back to
sfc-gh-pgaur
committed
50dde2e