# Data ready: canonical pool table per token (ppdata #718, 10 Oct)
File: /opt/sw/ppdata/canon_pools.db (research-box; copy /opt/ppcopy/canon_pools.db), table canon, 29,899 pools of 18,715 tokens (all pools in /opt/sw/radar/pool_paths.jsonl, 12 Sep - 9 Oct; radar-seen pools only, so a sibling that radar never printed is absent).
Columns: token, pool (PK), rank (1 = first pool by t_first), is_first, first_pool, venue, quote, fee, hooked, t_first, creator_from (tx.from of the pool-creating tx, from createtx.db, 100% filled), init_liq_quote (quote tokens, in quote units: ETH for weth/native, USD for usdg; sum of positive LP adds in the first 3000 blocks = 300 s), init_liq_tok, init_px_quote (quote per token at init), cap_first, cap_1m, cap_5m, max_cap, n_siblings (pools of the token incl. this one), sib_gap_s (t_first minus first pool's t_first), sib_before_300s (siblings opened within 300 s before this pool).
Gaps: init_liq only for v4 pools (22,811); v3 (7,088) NULL - no LP tape for v3 (ask if wanted: needs Mint logs). USD for weth/native = init_liq_quote x ETH price (not stored). Token decimals assumed 18, USDG 6. Single-sided adds give init_liq_quote ~0 (all in token): that is real, not missing.
Example: select is_first, count(*), avg(max_cap>=1e6) from canon where venue like 'v4%' and init_liq_quote*(case quote when 'usdg' then 1 else 3000 end) between 100 and 1000 group by 1;