Querying blocks

A worked example of pulling a block and its transactions from Chaingraph.

Query

query LatestBlock {
  block(limit: 1, order_by: { height: desc }) {
    hash
    height
    timestamp
    transaction_count
    size_bytes
    transactions(limit: 5) {
      hash
      size_bytes
      input_count
      output_count
    }
  }
}

Send

curl https://solidpeer.io/chaingraph/mainnet/<token> \
  -H "content-type: application/json" \
  -d '{"query":"query LatestBlock { block(limit:1, order_by:{height:desc}) { hash height timestamp transaction_count size_bytes transactions(limit:5) { hash size_bytes input_count output_count } } }"}'

Response shape

{
  "data": {
    "block": [
      {
        "hash": "00000000...",
        "height": 882357,
        "timestamp": 1715300000,
        "transaction_count": 64,
        "size_bytes": 51842,
        "transactions": [
          { "hash": "abc...", "size_bytes": 226, "input_count": 1, "output_count": 2 },
          ...
        ]
      }
    ]
  }
}

Filtering

Every list-shaped root field accepts a where: argument. Some examples:

# Blocks in a height range
block(where: { height: { _gte: 800000, _lte: 800100 } }) { hash height }

# Blocks at-or-after a wall-clock timestamp
block(where: { timestamp: { _gte: 1709000000 } }) { hash height timestamp }

# A specific block by hash (singleton form)
block_by_pk(hash: "<64-hex>") { height timestamp transaction_count }

Pagination

Use limit: + offset: for naive pagination, or cursor-style by carrying the last-seen height in subsequent where: filters. Cursor-style is faster on deep pagination — Hasura's offset does a sequential scan past skipped rows.

Cost

Computed from query complexity (depth × per-field weight × multiplier). The audit-log entry for each call shows the score; per-cycle usage in the dashboard rolls it up. Mainnet base cost is 300 CC; the score scales with how much work the query asks Hasura / Postgres to do.

If a customer's query exceeds their account balance mid-execution, the gateway returns 429 balance before the response is delivered.