# dataeng3 status 10 Oct ~11:30 UTC
## #207 shared tx.from
- NEW /opt/ppcopy/txfrom_all.db (research-box, rebuilt by /opt/sw/dataeng/txfrom_all.py, idempotent): table txf(tx PK, block, tx_from, src). 578,165 txs merged from v3tx (early v3 swaps), v2ramp_tx (111 V2RAMP pairs, 5,013 swaps, all with tx_from), txshared, ppdata txfrom.db (325k swaps / 3,884 pools, blocks 68.35M-82.60M).
- Query: sqlite3 /opt/ppcopy/txfrom_all.db "select tx_from from txf where tx='0x..'"; v2ramp swaps: /opt/sw/dataeng/v2ramp_tx.db swaps(pair,block,li,tx,side,weth,tx_from,recipient) + pairs.
- Gap: the full "first 30 min of every pool, all venues" (27k pools) is NOT covered; ppdata txfrom only 3.9k pools. Full coverage is millions of getTransactionByHash; not started (RPC cost). Ask the analyst which pool set matters and I extend txfrom_all for that set.
## #345 Proxima
- Probe 2: 40 newest of the 182 detected bundle launches (bundle_fund.db bundle_launch): token Transfers from bundle wallets to non-bundle addresses in the 7000 blocks after launch: 39/40 pools show ZERO, 1 pool one transfer. So the same-block bundles we already detect do NOT do Proxima-style delayed hand-offs; with probe 1 (no 5% fee leg on sells) this set is not Proxima. Proxima launches, if present, are in pools our bundle detector does not flag (script /opt/crew/work/dataeng3/px3.py, 40 paid calls).
- Next if wanted: scan all pools for "block-0 buyers who forward tokens to fresh wallets within 5-60 min" (needs Transfer logs per token, ~1 call per pool).
## #198
- /opt/ppcopy/v2swaps.db already exists (built by another role): tokmeta 2,666, pairs_sel 64 dec9 + 200 ctl, swaps 33k rows but tx_from filled for only 6.6k so far. I did not duplicate; not mine to finish.
## #126
- ppdata r6hook still running (hooked pools); not mine.