Regression in JDBC 0.9.8 when reading Array(Tuple(...)) with named tuple elements.

Start by reproducing the two SQL queries with clickhouse-jdbc 0.9.8, then compare the returned Array(Tuple(...)) value with JDBC 0.6.0 and the unnamed-tuple case. Check whether the driver or IntelliJ Database Viewer controls rendering; done means named tuple arrays expose their contents consistently with the expected cases.

Description

Regression in JDBC 0.9.8 when reading Array(Tuple(...)) with named tuple elements.

With the official ClickHouse JDBC driver 0.9.8, IntelliJ IDEA displays Array(Tuple(id UUID)) values as:
com.clickhouse.jdbc.types.Array@54c3b772
instead of the array contents.

The same query works correctly with JDBC 0.6.0.
The issue only affects named tuples. Arrays of unnamed tuples are displayed correctly.

Steps to reproduce

1. Connect to ClickHouse using clickhouse-jdbc:0.9.8.
2. Execute SELECT [('550e8400-e29b-41d4-a716-446655440000')]; Result is displayed correctly.
3. Execute SELECT [('550e8400-e29b-41d4-a716-446655440000')::Tuple(id UUID)]; Result is displayed as com.clickhouse.jdbc.types.Array@...

Error Log or Exception StackTrace

No exception is thrown.

The value is rendered as:

com.clickhouse.jdbc.types.Array@54c3b772

Expected Behaviour

Array(Tuple(id UUID)) should be returned through JDBC in a way that clients can display its contents, the same as in JDBC 0.6.0 and the same as Array(Tuple(UUID))

Code Example

SELECT [('550e8400-e29b-41d4-a716-446655440000')];

SELECT [('550e8400-e29b-41d4-a716-446655440000')::Tuple(id UUID)];

Configuration

Client Configuration

Default JDBC configuration.

Environment

• Cloud
• Client version: clickhouse-jdbc 0.9.8
• Language version: N/A (observed in IntelliJ Database Viewer)
• OS: macOS
• IDE: IntelliJ IDEA 2026.2.0.1

ClickHouse Server

• ClickHouse Server version: 26.3.17.56
• ClickHouse Server non-default settings, if any: None
• CREATE TABLE statements for tables involved: Not required (reproducible with literal values only)
• Sample data for all these tables, use clickhouse-obfuscator if necessary

It is unclear whether this is a JDBC driver regression or an incompatibility with IntelliJ’s database viewer, but the behavior changed between JDBC 0.6.0 and 0.9.8 and is reproducible with the SQL above.