Incorrect list of columns for table containing underscore

System information:

• Operating system (distribution) and version: both Windows and Linux
• DBeaver version: 7.2.2 (issue present in previous versions too)
• Additional extensions: none

Connection specification:

• Database name and version: SAP HANA (both 2.0 and HANA Cloud)
• Driver name: maven:/com.sap.cloud.db.jdbc:ngdbc, versions 2.4, 2.5, 2.6
• Do you use tunnels or proxies (SSH, SOCKS, etc)? no

Describe the problem you’re observing:

A table with the name containing an underscore will display an incorrect list of columns, if there is another table with a letter instead of the underscore in the name.

Steps to reproduce, if exist:

1. Create two tables, ABC and A_C.
2. Double click on A_C, go to Properties, then Columns. You will see the columns from both tables, first the ones from ABC and then the ones from A_C.

The reason most likely is that underscore matches any letter in the SQL LIKE predicate: https://help.sap.com/viewer/4fe29514fd584807ac9f2a04f6754767/2.0.05/en-US/20fa17f375191014a4d8d8cbfddfe340.html

If this is the case, other databases might also be affected, as this is not HANA-specific - it is also specified in the ANSI SQL standard. In general, a check for equality should be done using the = operator instead, while LIKE should be used for pattern matching only.

Include any warning/errors/backtraces from the logs

[kai-morich]

the internal statement for DatabaseMetaData#getColumns contains

  WHERE 
     ( ($2) IS NULL  OR ($2) = '' OR C.SCHEMA_NAME LIKE ($2) ESCAPE '\')
     AND C.TABLE_NAME LIKE ($3) ESCAPE '\'
     AND C.COLUMN_NAME LIKE ($4) ESCAPE '\'

so \ is the escape character

[kai-morich]

Of course would be better if the JDBC driver would provide this option but this will take some time. Many JDBC drivers are incomplete so dbeaver has various places to compensate for missing properties like SQLDialect classes and driver plugin.xml