Jump to a popular page, or start typing.

    CONNECTIONS / Clickhouse

    Clickhouse + Datahash: event tables become ad signals.

    Event and customer data land in Clickhouse tables built for fast scans over huge volumes. Signals reads them on a schedule you set and delivers hashed events and audiences to every major ad platform, with no export.

    clickhouse :: live

    1. ● clickhouse :: signals live
    2. > SELECT * FROM events WHERE event_time > :cursor rows=118,204
    3. > read customer_attributes role=readonly
    4. hash sha256(email,phone)
    5. route meta · google · tiktok · snap
    6. ✓ 4 destinations synced · match 81%
    WHAT FLOWS THROUGH CLICKHOUSE

    The signal Clickhouse yields, and receives.

    Every kind of signal that moves between Clickhouse and the platforms, through one hashed, deduplicated route.

    What flows Every platform Offline conversions Lead conversions Custom audiences Lead generation Signals hash · dedupe· route Google Meta Snap TikTok LinkedIn OpenAI
    Offline conversions
    Clickhouse tables are typically ordered by a timestamp or event key, so Signals can read past a cursor on that column without a full table scan even at hundreds of millions of rows.
    Lead conversions
    Because Clickhouse organizes data into partitions, a run touches only the partitions holding new or updated rows, keeping load on the cluster low as event volume grows.
    Custom audiences
    Emails and phone numbers are SHA-256 hashed at ingestion, so no plaintext identifier from a Clickhouse table ever reaches a destination.
    Lead generation inbound
    The visual mapper aligns whatever column names a MergeTree or Distributed table uses to each platform's event fields, so clickstream or log-shaped data still becomes clean conversion signals.
    USE-CASE LISTING

    Pick your platform. Every use case for it, in one place.

    Every way Clickhouse data moves, grouped by platform and ranked by how many use cases Clickhouse supports.

    Google 5 Meta 4 Snap 4 TikTok 4 LinkedIn 3 OpenAI 1
    HOW IT CONNECTS

    From authorization to delivered signal.

    1. Connect

      Connect Clickhouse with a read-only user restricted to the tables you name, over an encrypted channel.

    2. Select

      Point Signals at the tables or views that hold conversion and event data.

    3. Map

      Map columns to events and identifiers in the visual mapper.

    4. Go live

      Verify sample rows in the live debugger and go live.

    FAQ

    Clickhouse questions, answered.

    Does Datahash write to my Clickhouse cluster?

    No. Signals connects with a read-only user restricted to the tables and views you name, so it can scan rows but never insert, update, or delete data. Clickhouse's own grants let you scope that account to specific databases, and you can revoke the credential at any time without affecting ingestion into the rest of your cluster.

    Is customer PII stored by Datahash?

    Identifiers are SHA-256 hashed at ingestion and no PII is stored on host servers in Datahash Cloud, so only hashed values ever reach a destination platform. If event data cannot leave your perimeter, a self-hosted Core deployment runs the same hashing step inside your own environment before anything is delivered.

    How often does the sync run against Clickhouse?

    On a schedule you choose, from every few minutes to daily batches. Because Clickhouse is built to scan very large tables quickly, incremental queries against a growing MergeTree table stay fast, so a high-volume event table can sync as often as a small reference table without straining the cluster.

    NEXT STEP

    See your Clickhouse data working in every ad platform.

    Ready to enable a use case, or still mapping what Clickhouse data could do? Our team helps you find the right place to start.