Arbitrum

Seven Parquet tables with 295 documented columns in total: blocks (22), transactions (40), calls (65), events (63), transfers (24), dex_trades (54) and miner_rewards (27). Each purchase covers the window you choose, through to yesterday, refreshed daily. The Full tier starts at the Nitro upgrade on 31 August 2022 (block 22,207,818) for six tables; miner_rewards starts on 8 May 2026 (block 460,491,246).

To block 22,207,818 on 31 August 2022, the day the Nitro upgrade went live. Arbitrum Classic, the chain's history from May 2021 until Nitro, is not in our index. User transfers and trades start at about 16:08 UTC that day.

The largest tier by a wide margin, and calls and events dominate it; both carry several rows per transaction, where transfers and dex_trades carry one per movement. We size it per table before delivery and send you the exact figures with the manifest, so you can plan storage before downloading. If calls and events are more than you need, we can quote the other five tables on their own.

On Transaction_Hash, which appears in every table except blocks and miner_rewards; those join on Block_Number. Within a transaction, Call_Index orders the call frames, Log_Index orders the logs, Transfer_Index orders the transfers and Trade_Index orders the swaps. A log points back to the frame that emitted it through Call_Index.

Balance updates. The chain-wide balance_updates table exists in our store but is not part of this bundle, because it carries several rows per transfer and would dominate the size. If you need it, ask and we will quote it alongside this one.

As parallel arrays rather than one JSON column: element i of Arguments_Type_Name, Arguments_Type_Type and the Arguments_Value_* columns all describe argument i. Each value sits in the column matching its ABI type, addresses and bytes in Arguments_Value_Bytes as hex, small integers in Arguments_Value_UInt or Arguments_Value_Int, and every value also rendered in Arguments_Value_String. The Arguments_Path_* arrays give the member path for struct and array arguments. In calls, the Returns_* columns follow the same layout for return values.

Arbitrum closes a block about every 250 milliseconds, and sometimes several per second, so the chain passed 500 million blocks in 2026. Files are still cut every 50 blocks, which here means a file spans seconds; a one-month window is tens of thousands of files per table. The manifest lists them all, and the files sort by block number.

The fee columns carry what the execution layer records for the transaction: gas used, effective gas price, base fee and priority fee, in ETH and USD. The separate L1 posting cost that Arbitrum folds into its gas price is not broken out as its own column. If you need it, ask and we will scope it.

No. Arbitrum One has a single sequencer and produces no uncle blocks, so Block_UnclesCount is 0 on every block, Block_UncleHash is the empty-list hash, and Reward_Uncle, Uncle_Block_Number and Uncle_Index in miner_rewards are 0 throughout. The columns exist because the schema is shared with chains that do have uncles; they are kept so a query written against one EVM chain runs unchanged here.

Every venue we decode on Arbitrum One, in one table, identified by Trade_Dex_ProtocolFamily and Trade_Dex_ProtocolName on each row. In the sample window the families present were Uniswap and PancakeSwap. Filter the family column to keep or drop a venue.

Only currencies with an hourly price series are valued, and the dataset says so rather than filling the gap: a USD value of 0 means either a zero-value row or no price, and the column cannot tell the two apart. DEX trades are priced from the safer side of the swap, so most swaps with a stablecoin or ETH on one side carry a value; transfers of long-tail and spam tokens mostly do not.

Yes, throughout, and the export applies no filter of its own. TransactionStatus_Success carries the transaction's outcome on every table that has it, and dex_trades additionally carries Trade_Success, because a single swap can fail inside a transaction that succeeded. Spam tokens, zero-value rows and lookalike contracts are all kept too; the columns let you drop what you do not want.

Filter Call_To in calls and Log_SmartContract in events to the contract, lower case with 0x. For token flow, filter Transfer_Currency_SmartContract in transfers. Join on contract addresses rather than symbols or names; both repeat, and lookalike tokens copy them deliberately.

Amounts are decimal-adjusted decimal strings, carrying exactly as many places as the token has decimals; multiply by ten to the power of the decimals for the raw on-chain integer. Casting to float loses precision on large values. Boolean flags are the strings true and false; only the Arguments_Value_Bool and Returns_Value_Bool arrays use 0 and 1, one entry per decoded argument.

One line in DuckDB: SELECT * FROM read_parquet('transfers/*.parquet') LIMIT 10. Pandas works too: pandas.read_parquet(path). The sample files open without credentials.

As flat Parquet named by block range, 50 blocks per file and sorted ascending, one prefix per table, with a JSON manifest listing every file and its sha256. We prepare the files within 7 business days of purchase and email you the location. You choose at checkout where to collect them: a Requester Pays S3 bucket, downloaded with your own AWS account, or a Cloudflare R2 bucket with a read-only API key and no egress fees. Files stay available for 7 days from delivery. Enterprise orders can be delivered into your own S3, GCS or R2 bucket. Refreshed daily with T+1 latency, so a purchase made today includes everything to yesterday.

Yes. Each table links to a public sample file with real records, and the sample bucket holds full Parquet files you can load directly. No email address required.