Mixed-case behavior based on connectors

From IBM® watsonx.data version 2.0.0, a new feature is available to switch between case-sensitive and case-insensitive behavior in Presto (Java) by using a mixed-case feature flag. This feature is not applicable for Presto (C++) engine. The mixed-case feature flag is set to OFF in Presto (Java) by default. The flag can be set to ON or OFF as required during deployment of the Presto (Java) engine. It is advised not to toggle between ON and OFF configurations after the deployment, as it may result in inconsistent system behavior.

Note: For more information on mixed-case feature flag behavior, supported SQL statements and supported data types matrices, see Support content.

watsonx.data Developer edition

watsonx.data on IBM Software Hub

Mixed-case feature flag: ON

You can enable mixed-case identifiers (schema, table & column names) per catalog using a global API configuration in Presto (Java). When you set the global property enable-mixed-case-support=true, the system automatically enables a catalog-level property case-sensitive-name-matching=true for all eligible catalogs.

The following catalogs support the case-sensitive-name-matching property:
  • HANA
  • MySQL
  • Oracle
  • PostgreSQL
  • Amazon Redshift
  • SingleStore
  • SQL Server
  • IBM Db2
  • IBM Data Virtualization Manager
  • IBM Db2fori
  • IBM Informix
  • IBM Netezza
  • Apache Phoenix
  • Snowflake
  • Teradata
  • BigQuery
  • Greenplum
  • Apache Derby
  • MariaDB
  • Apache Cassandra
  • ClickHouse
  • Apache Druid
  • Elasticsearch
  • Apache Kafka
  • MongoDB
  • Apache Pinot
  • Redis
  • Prometheus
Note: Presto defaults the case-sensitive-name-matching configuration to false for all eligible catalogs.

Mixed-case feature flag: OFF

When case-sensitive-name-matching=false (the default setting), Presto (Java) converts identifiers to lowercase. This is the default behavior in Presto (Java).