版本：3.7.1

06 Bitcoin Data Verification

To verify that Bitcoin block and transaction data is complete in ClickHouse, you need to ensure:

No missing blocks
No missing transactions
No duplicate block heights
Transaction counts match block metadata

1. block height continuity

Run this to detect missing block numbers:

WITH seq AS (
    SELECT number
    FROM numbers(
        toUInt64(
            ifNull((SELECT max(number) FROM blocks_fat), 0) + 1
        )
    )
)
SELECT seq.number AS missing_block_number
FROM seq
LEFT ANTI JOIN blocks_fat AS b ON seq.number = b.number

If this query returns rows, those block numbers are missing.

2. Check for duplicate blocks

Check for duplicate block numbers (heights):

SELECT number, count() AS cnt
FROM blocks_fat
GROUP BY number
HAVING cnt > 1

You should see zero rows returned.

if the count of one block is bigger than 1, please use final command to deduplicate the same rows

OPTIMIZE TABLE blocks_fat FINAL;

3.Transactions Missing

To detect missing transactions from the transactions_fat table that are listed in the blocks_fat.transactions array, and process it partition by partition

To list all the partition in the blocks_fat table

SELECT DISTINCT toYYYYMM(timestamp_month) AS part
FROM blocks_fat
ORDER BY part ASC

To find the transactions which are missing from transactions_fat table partition by partition

WITH flattened AS (
    SELECT
        hash AS block_hash,
        arrayJoin(transactions) AS tx_hash
    FROM blocks_fat
    WHERE toYYYYMM(timestamp_month) = 201304
)
SELECT
    flattened.block_hash,
    flattened.tx_hash,
    t.hash
FROM flattened
LEFT JOIN (
    SELECT hash
    FROM transactions_fat
    WHERE toYYYYMM(block_timestamp_month) = 201304
) AS t
ON flattened.tx_hash = t.hash
WHERE t.hash != flattened.tx_hash;

The missing transaction examples

Block: 364773  Missing TX: 9fdbcf0ef9d8d00f66e47917f67cc5d78aec1ac786e2abb8d2facb4e4790aad6
Block: 364784  Missing TX: 9c667c64fcbb484b44dcce638f69130bbf1a4dd0fbb4423f58ceff92af4219ec
Block: 367885  Missing TX: 30b3b19b4d14fae79b5d55516e93f7399e7eccd87403b8dc048ea4f49130595a
Block: 367906  Missing TX: dd6067e71c04cb62f8e5aa52ecc99b01ffcd551a52727d046a2fabb14eb39b4d
Block: 367904  Missing TX: 740ac533882221099e7202bbdafbb99ec589c6e74fd2fe7ca1274b46ea4f0a96
Block: 367899  Missing TX: f2e197a6d8d088b13afd0f99d4027da36a9413b9f3d7730ba5278132ebc950a7
Block: 367886  Missing TX: cf1032c2213e6faea04f1813aa6890e7f588bb378cb98e7425aec83c11d4457c
Block: 367877  Missing TX: 52539a56b1eb890504b775171923430f0355eb836a57134ba598170a2f8980c1
Block: 367911  Missing TX: 5f4d2593c859833db2e2d25c672a46e98f7f8564b991af9642a8b37e88af62bc
Block: 367897  Missing TX: 8dabbf51f78c1e7286866af1de403118c5ddbe57ca93b54859245916d2bf1063
Block: 367891  Missing TX: c9fe64681c9a12795586a3ae7c5e94b585032f67847c7f9c42e1b979a1e2959b
Block: 367906  Missing TX: dd6067e71c04cb62f8e5aa52ecc99b01ffcd551a52727d046a2fabb14eb39b4d
Block: 367904  Missing TX: 740ac533882221099e7202bbdafbb99ec589c6e74fd2fe7ca1274b46ea4f0a96
Block: 367899  Missing TX: f2e197a6d8d088b13afd0f99d4027da36a9413b9f3d7730ba5278132ebc950a7
Block: 367885  Missing TX: 30b3b19b4d14fae79b5d55516e93f7399e7eccd87403b8dc048ea4f49130595a
Block: 367906  Missing TX: dd6067e71c04cb62f8e5aa52ecc99b01ffcd551a52727d046a2fabb14eb39b4d
Block: 367904  Missing TX: 740ac533882221099e7202bbdafbb99ec589c6e74fd2fe7ca1274b46ea4f0a96
Block: 367899  Missing TX: f2e197a6d8d088b13afd0f99d4027da36a9413b9f3d7730ba5278132ebc950a7
Block: 367886  Missing TX: cf1032c2213e6faea04f1813aa6890e7f588bb378cb98e7425aec83c11d4457c
Block: 367877  Missing TX: 52539a56b1eb890504b775171923430f0355eb836a57134ba598170a2f8980c1
Block: 367911  Missing TX: 5f4d2593c859833db2e2d25c672a46e98f7f8564b991af9642a8b37e88af62bc
Block: 367897  Missing TX: 8dabbf51f78c1e7286866af1de403118c5ddbe57ca93b54859245916d2bf1063
Block: 367891  Missing TX: c9fe64681c9a12795586a3ae7c5e94b585032f67847c7f9c42e1b979a1e2959b
Block: 367885  Missing TX: 30b3b19b4d14fae79b5d55516e93f7399e7eccd87403b8dc048ea4f49130595a
Block: 367911  Missing TX: 5f4d2593c859833db2e2d25c672a46e98f7f8564b991af9642a8b37e88af62bc
Block: 367877  Missing TX: 52539a56b1eb890504b775171923430f0355eb836a57134ba598170a2f8980c1
Block: 367897  Missing TX: 8dabbf51f78c1e7286866af1de403118c5ddbe57ca93b54859245916d2bf1063
Block: 367891  Missing TX: c9fe64681c9a12795586a3ae7c5e94b585032f67847c7f9c42e1b979a1e2959b
Block: 367886  Missing TX: cf1032c2213e6faea04f1813aa6890e7f588bb378cb98e7425aec83c11d4457c

If you find there are transactions missed, please double check whether it contained in the bitcoind

bitcoin-cli getrawtransaction <txid> true

Apache Kafka has several key limitations and configurable constraints that impact how much data can be produced, buffered, and consumed. Here's a clear breakdown of the main types of limitations:

Message Size Limitations
Buffering & Memory Limits
Log & Retention Limits
Consumer Fetch & Processing Limits
Connection & Network Limits

To enlarge your Kafka topic transactions to support messages up to 90MB, you must:

Update the topic config (max.message.bytes)
Update the broker config (message.max.bytes and replica.fetch.max.bytes)
Update your producer config (max.request.size, buffer.memory)

It works after finish the first step to update the transactions topic config.

kafka-configs.sh --bootstrap-server localhost:9092 \
  --entity-type topics \
  --entity-name transactions \
  --alter \
  --add-config max.message.bytes=94371840

kafka-configs.sh --bootstrap-server localhost:9092   --entity-type topics --entity-name transactions   --describe

Dynamic configs for topic transactions are:
  cleanup.policy=delete sensitive=false synonyms={DYNAMIC_TOPIC_CONFIG:cleanup.policy=delete, DEFAULT_CONFIG:log.cleanup.policy=delete}
  max.message.bytes=94371840 sensitive=false synonyms={DYNAMIC_TOPIC_CONFIG:max.message.bytes=94371840, DEFAULT_CONFIG:message.max.bytes=1048588}
  retention.bytes=1073741824 sensitive=false synonyms={DYNAMIC_TOPIC_CONFIG:retention.bytes=1073741824, DEFAULT_CONFIG:log.retention.bytes=-1}
  retention.ms=86400000 sensitive=false synonyms={DYNAMIC_TOPIC_CONFIG:retention.ms=86400000}

4.Transactions Missing Reverse

To find transactions listed in transactions_fat that are not present in blocks_fat.transactions, partition-by-partition (month-by-month)

WITH
    transactions_fat_partitioned AS
    (
        SELECT
            hash,
            block_hash
        FROM transactions_fat
        WHERE toYYYYMM(block_timestamp_month) = 201304
    ),
    blocks_fat_partitioned AS
    (
        SELECT
            hash AS block_hash,
            arrayJoin(transactions) AS tx_hash
        FROM blocks_fat
        WHERE toYYYYMM(timestamp_month) = 201304
    )
SELECT
    t.hash AS missing_transaction_hash,
    t.block_hash AS expected_block_hash,
    b.tx_hash
FROM transactions_fat_partitioned AS t
LEFT JOIN blocks_fat_partitioned AS b ON (t.hash = b.tx_hash) AND (t.block_hash = b.block_hash)
WHERE b.tx_hash != t.hash

4. duplicate transactions

SELECT hash, count() AS cnt
FROM transactions_fat
GROUP BY hash
HAVING cnt > 1

Again, this should return no rows.

5. Python to Loop Partition by Partition

Here is a Python script that:

Connects to ClickHouse via clickhouse-connect.
Iterates over partition months (you define the range).
Runs a query per partition to find transactions listed in blocks_fat but missing from transactions_fat.

from clickhouse_connect import get_client
from datetime import datetime
from dateutil.relativedelta import relativedelta

# === CONFIGURATION ===
CLICKHOUSE_HOST = 'localhost'
CLICKHOUSE_PORT = 8123
CLICKHOUSE_USER = 'default'
CLICKHOUSE_PASSWORD = 'password'
DATABASE = 'bitcoin'

START_MONTH = '2010-01'  # yyyy-mm
END_MONTH   = '2013-11'

# === INIT CLIENT ===
client = get_client(
    host=CLICKHOUSE_HOST,
    port=CLICKHOUSE_PORT,
    username=CLICKHOUSE_USER,
    password=CLICKHOUSE_PASSWORD,
    database=DATABASE,
)

# === GENERATE PARTITIONS ===
def generate_partitions(start, end):
    partitions = []
    start_date = datetime.strptime(start, '%Y-%m')
    end_date = datetime.strptime(end, '%Y-%m')
    current = start_date
    while current <= end_date:
        partitions.append(current.strftime('%Y%m'))
        current += relativedelta(months=1)
    return partitions

# === MAIN LOOP ===
partitions = generate_partitions(START_MONTH, END_MONTH)

for partition in partitions:
    print(f'\n🧪 Querying Partition: {partition}')
    query = f"""
    WITH flattened AS (
        SELECT
            hash AS block_hash,
            tx_hash
        FROM bitcoin.blocks_fat
        ARRAY JOIN transactions AS tx_hash
        WHERE toYYYYMM(timestamp_month) = {partition}
    )
    SELECT
        flattened.block_hash,
        flattened.tx_hash
    FROM flattened
    LEFT JOIN (
        SELECT hash
        FROM bitcoin.transactions_fat
        WHERE toYYYYMM(block_timestamp_month) = {partition}
    ) AS txs
    ON flattened.tx_hash = txs.hash
    WHERE txs.hash != flattened.tx_hash
    """
    result = client.query(query)
    rows = result.result_rows
    if rows:
        print(f"⚠️  Missing transactions in partition {partition}: {len(rows)}")
        for row in rows:
            print(f"    Block: {row[0]}, Missing TX: {row[1]}")
    else:
        print(f"✅ All transactions found in partition {partition}")

pip install clickhouse-connect python-dateutil

1. block height continuity​

2. Check for duplicate blocks​

3.Transactions Missing​

4.Transactions Missing Reverse​

4. duplicate transactions​

5. Python to Loop Partition by Partition​