In this post, we’ll look at updating and deleting rows with ClickHouse. It’s the second of two parts.

In the first part of this post, we described the high-level overview of implementing incremental refresh on a ClickHouse table as an alternative support for UPDATE/DELETE. In this part, we will show you the actual steps and sample code.

Prepare Changelog Table

First, we create the changelog table below. This can be stored on any other MySQL instance separate from the source of our analytics table. When we run the change capture script, it will record the data on this table that we can consume later with the incremental refresh script:

CREATE TABLE `clickhouse_changelog` (
  `db` varchar(255) NOT NULL …

In this post, we’ll look at updating and deleting rows with ClickHouse. It’s the first of two parts.

Update: Part 2 of this post is here.

ClickHouse is fast – blazing fast! It’s quite easy to pick up, and with ProxySQL integrating with existing applications already using MySQL, it’s way less complex than using other analytics options. However, ClickHouse does not support UPDATE/DELETE (yet). That entry barrier can easily dissuade potential users despite the good things I mentioned.

If there is a will, there is a way! We have so far taken advantage of the new feature that supports more granular partitioning strategy (by week, by day or something else). With more …