Question

Imagine I have a query like

SELECT * from table1 a, table2 b where (WHATEVER)

Maybe both tables have the same column name. So I though it would be nice to access the data via

resultSet.getString("a.columnName");
resultSet.getString("b.columnName");

But this backfires on me and I get nothing. I read the API, but they don’t really talk about this case. Is such a feature vendor dependent?

Answer 1

JDBC will simply name the columns by what is specified in the query – it doesn’t know about table names etc.

You have two options:

Option 1: Name the columns differently in the query, ie

SELECT
    a.columnName as columnNameA,
    b.columnName as columnNameB,
    ...
from table1 a, table2 b where (WHATEVER)

then in your java code refer to the column aliases:

resultSet.getString("columnNameA");
resultSet.getString("columnNameB");

Option 2: Refer to the column position in your call to the JDBC API:

resultSet.getString(1);
resultSet.getString(2);

Note that the JDBC API uses one-based indexes – ie they count from 1 (not from 0 like java indexes), so use 1 for the first column, 2 for the second column, etc

I would recommend option 1, because it’s safer to refer to named columns: Someone may change the order of the columns in the query and it would silently break your code (you would be accessing the wrong column but would not know), but if they change the columns names, you’ll at least get a “no such column” exception at runtime.