ByteHouse 1.1.0.0 Release, September 2021

In version 1.1.0.0 of ByteHouse, we are pleased to announce the introduction of bucket tables, auto-resume and suspend configurations for virtual warehouses, Google Cloud support, and many more new features and improvements.

📘

Get in touch

If you are interested to try out our preview features, get in touch with us by creating a ticket on the support portal.

New features

Bucket tables

Similar to partitions, bucket tables provide you with another method for decomposing table data sets into more manageable parts.

Unlike partitions that require a logical criterion (e.g. split by month, or day), bucket tables are split based on a hash function that is applied to a column, and the number of buckets is specified when you create a table.

Advantages:

  • ByteHouse will only access applicable buckets during a query, reducing the amount of data that is needed to be loaded.
  • Co-located data can significantly improve specific join, aggregate, and lookup performance. (e.g. two tables using the same cluster key to generate buckets, then execute co-locate join/group by on the micro-partitions to improve performance)
  • For point queries, it could specify a bucket to get query results by using cluster key and reach higher QPS

Auto-resume and suspend virtual warehouses

Auto-resume and suspend virtual warehouses on ByteHouse. To enable this option, you can configure auto-resume and suspend when creating a virtual warehouse.

Advantages:

  • Reduce your running costs, by running the virtual warehouse only when it is needed. This is especially for workloads that have fluctuating usage patterns.

Adding Confluent Cloud new data source

ByteHouse now supports having Confluent Cloud as a streaming data source.

Column Expressions [In-Preview]

ByteHouse now supports Column Expressions when performing data loadings. This feature allows you to perform some simple pre-processing while performing a data loading job.

This is a preview feature, please contact support for more info on this feature.

Supported providers and regions: Google Cloud [In-Preview]

ByteHouse now supports deployment on Google Cloud (Singapore).

This is a preview feature, please contact support for more info on this feature.

JDBC driver [In-Preview]

ByteHouse has published the first version of our JDBC driver.

This is a preview feature, please contact support for more info on this feature.

Improvements

User Interface

  • Improved accessibility for all sub-pages
  • Add tooltip helper if virtual sizes are not suitable
  • Prettify and localize ByteHouse Errors
  • Add prompt when new SQL file is uploaded
  • Prevent tour prompt after skipping
  • Translation refining
  • Add tooltips about incremental loading for AWS S3 job

Performance

  • Support Create Table AS SELECT * Syntax
  • Speed up compaction job

Fixed bugs

  • SQL Parser: YEAR and DATE is backtick-ed and perceived as column name in CNCH
  • Fixed order by / decimal order by
  • Unblock submit executions using security rules
  • Bucket Table DDL Cluster by () now works with parenthesis
  • Enable using DB after creating a database
  • Fixed JSON format in operator metrics