We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Olga Silyutina- ClickHouse Applications in Data Analytics | PyData Yerevan July 2022 Meetup
Explore the capabilities of ClickHouse in data analytics, including OLAP scenarios, custom functions for analytics and machine learning, and optimized data storage and retrieval methods.
- ClickHouse is a column-oriented database mainly used for online analytical processing (OLAP) scenarios.
- ClickHouse is not suitable for online transaction processing (OLTP) due to limitations in storing data and missing some data during insertion.
- Custom functions in ClickHouse can be used for analytics and machine learning (ML) processes.
- Multi-function in ClickHouse is a compact and easier-to-read version of the “CASE WHEN” function in SQL.
- Engine, such as Merge Tree, is used in ClickHouse to store data and make it more efficient.
- Partitions in ClickHouse are like folders that store data in a specific format, making it easier to retrieve and filter data.
- Materialized views in ClickHouse can be used to create real-time aggregates based on select queries and can be faster than basic views.
- Array functions in ClickHouse, such as group arrays and array enumerate, can be used for ranking and creating aggregates.
- Low cardinality strings in ClickHouse can compress strings and make them more efficient.
- ClickHouse supports replication and sharding, making it possible to store data on multiple machines and speed up processing.
- ClickHouse has a unique exact function that calculates the unique amount of ad IDs in a table.
- Approximate calculations in ClickHouse can be used to make calculations faster and more efficient.
- Sample by index in ClickHouse can be used to get samples of data based on a specific column.
- OLAP is a data discovery process that requires low latency and frequent queries.
- ClickHouse has a lot of integrational engines, such as Kafka, MySQL, and JDBC.
- Unique functions in ClickHouse can be used to calculate unique amounts of data in a table.
- Materialized views can be used to transform data in a specific format and make it easier to retrieve and filter.
- ClickHouse compared to relational databases has multiple indexes that can be used in different ways.