prediction_markets.trades

The prediction_markets.trades table is the cross-venue trade table. Grain: one row per taker fill, including Polymarket Combos (multi-leg parlays; is_parlay = TRUE in the markets dimension). Each match appears once, so SUM(amount_usd) is venue volume without any filter. Polymarket maker legs stay in the venue table (polymarket_polygon.market_trades, is_taker_side = FALSE). Three columns describe every fill:

price is always the probability of Yes in [0, 1], comparable across venues and sides.
outcome_side is the side of the market the traded contract belongs to (yes or no).
taker_side is whether the taker bought or sold that contract (BUY or SELL).

taker_fill_price is the price of the contract actually traded (price for a Yes contract, 1 - price for a No contract), so taker_fill_price × num_contracts = amount_usd on every row.

Kalshi nets positions and reports every fill as a buy of the side the taker ends up long, so taker_side is always BUY on Kalshi and closing a Yes position appears as a BUY with outcome_side = 'no'. Kalshi also reports one row per taker-maker match, so row counts are not comparable across venues; sums are.

Table Schema

Column	Type	Description
`venue`	`VARCHAR`	`polymarket` or `kalshi`
`executed_at`	`TIMESTAMP`	When the trade executed
`block_month`	`DATE`	First day of the calendar month of `executed_at` (partition key)
`trade_id`	`VARCHAR`	Trade identifier, unique within a venue. On Polymarket `tx_hash:evt_index`
`market_id`	`VARCHAR`	Market traded: Polymarket `condition_id` (combo condition ID for Combos), Kalshi ticker. Joins `prediction_markets.markets` on `(venue, market_id)`
`event_id`	`VARCHAR`	Parent event of the market. Always populated on Kalshi; on Polymarket only for markets in a multi-outcome (neg-risk) event, NULL for Combos
`outcome_id`	`VARCHAR`	Identifier of the side traded: the token on Polymarket, the market itself on Kalshi. Keep it as text; Polymarket token IDs do not fit a numeric type. To label the side, join `prediction_markets.markets` and pick `yes_outcome_name` or `no_outcome_name` using `outcome_side`
`outcome_side`	`VARCHAR`	Side of the market the contract in `outcome_id` belongs to: `yes` (the side `price` refers to) or `no`. Same meaning on both venues
`taker_side`	`VARCHAR`	Whether the taker bought or sold the contract in `outcome_id`: `BUY` or `SELL`. Always `BUY` on Kalshi
`price`	`DOUBLE`	Probability of Yes at which the trade executed, in `[0, 1]`
`taker_fill_price`	`DOUBLE`	Price per contract of the contract traded: `price` when `outcome_side = 'yes'`, `1 - price` when `'no'`. Times `num_contracts` it equals `amount_usd`
`num_contracts`	`DOUBLE`	Contracts traded. One contract pays $1 if its side wins
`amount_usd`	`DOUBLE`	Cash that changed hands, in USD: the buyer’s cost and the seller’s proceeds. Use it for volume
`fee_usd`	`DOUBLE`	Taker fee in USD. Reported on-chain for Polymarket; estimated from Kalshi’s published fee formula for Kalshi, NULL where the series has no fee terms
`trader_address`	`VARCHAR`	The taker’s wallet (0x-hex). Polymarket only; NULL on Kalshi
`tx_hash`	`VARCHAR`	Transaction hash (0x-hex). Polymarket only; NULL on Kalshi
`_updated_at`	`TIMESTAMP`	When the trade was last written by the pipeline

Table sample

Query performance

block_month is the partition key. Always include a block_month or executed_at filter.

-- ✅ Good: time-bounded
SELECT * FROM prediction_markets.trades
WHERE block_month >= DATE '2026-05-01'
  AND venue = 'kalshi'

Example queries

-- Top markets by USD volume across both venues (last 7 days)
SELECT
  t.venue,
  t.market_id,
  m.question,
  COUNT(*) AS num_trades,
  SUM(t.amount_usd) AS total_volume_usd
FROM prediction_markets.trades t
JOIN prediction_markets.markets m
  ON m.venue = t.venue
  AND m.market_id = t.market_id
WHERE t.block_month >= DATE_TRUNC('month', CURRENT_DATE) - INTERVAL '1' MONTH
  AND t.executed_at >= NOW() - INTERVAL '7' DAY
GROUP BY 1, 2, 3
ORDER BY 5 DESC
LIMIT 20

-- Taker flow by side over the last 24 hours
SELECT
  venue,
  outcome_side,
  taker_side,
  SUM(amount_usd) AS volume_usd,
  SUM(num_contracts) AS contracts
FROM prediction_markets.trades
WHERE block_month >= DATE_TRUNC('month', CURRENT_DATE) - INTERVAL '1' MONTH
  AND executed_at >= NOW() - INTERVAL '1' DAY
GROUP BY 1, 2, 3
ORDER BY 1, 2, 3

Curated Data

Solana

Hyperliquid

Non-EVM Networks

Protocols

EVM Networks

Partner Data

Dune Index

prediction_markets.trades

Table Schema

Table sample

Query performance

Example queries

​Table Schema

​Table sample

​Query performance

​Example queries

Table Schema

Table sample

Query performance

Example queries