[CALCITE-835] Unicode character seems to be handled incorrectly in Avatica

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: avatica-1.8.0
• Component/s: avatica
• Assignee: Josh Elser
• Reporter: Holman Lan
• Created: 10/Aug/15 16:04
• Updated: 29/Apr/19 17:14
• Resolved: 30/Mar/16 03:49

Description

This was discovered with Apache Phoenix (4.4 and 4.5) while using the thin JDBC client to connect to Phoenix Query Server.

This can be reproduced using a CREATE TABLE statement with non-ASCII characters in column name:

create table colUnicode ("НомерТелефона" integer not null primary key, col2 varchar)

When executing the above statement using the thin client and retrieving the table using either the thin or the thick client the column name shows up as �оме�Теле�она. Executing the same CREATE table statement using the thick client seems to work fine, the column name shows up correctly when using either the thin or the thick client to retrieve the table.

The same behavior can also be observed when when inserting non-ASCII data into a varchar column in a table.

When using WireShark to observe the network traffic, we can see the non-ASCII character sent from the thin client to the server has been encoded into UTF-8.