使用Oracle SQL Developer中的表达式对多个列进行别名

Question

In Oracle SQL Developer, is it possible to alias multiple column names as part of a SELECT statement using an expression (as opposed to manually specifying the aliases for each column)?

Specifically, I have a mapping table that stores the task-relevant subset of columns from a large data table. Each entry in the mapping table ties a data table column name to a human readable description. I want to select the data table columns listed in the mapping table and display them with the mapping table descriptions as the column headers, but WITHOUT manually typing in the column names and their human-readable aliases one-by-one. Is this possible?

The closest I've found to an answer online is this SO question which suggests what I want to do is NOT possible: Oracle rename columns from select automatically?

But, that question is from 2010. I'm hoping the situation has changed. Thank you for your help.

Answer 1

This still cannot be done with 100% native SQL. These overly-dynamic situations are usually best avoided; a little extra typing is generally better than adding complicated code.

If you truly have an exceptional case and are willing to pay the price there is a way to do this. It doesn't use 100% natural SQL, but it could be considered "pure" SQL since it uses the Oracle Data Cartridge framework to extend the database.

You can use my open source project Method4 to run dynamic SQL in SQL. Follow the Github steps to download and install the objects. The code is painfully complicated but luckily you won't need to understand most of it. Only the simple changes below are necessary to get started on customizing column names.

Method4 Changes

Create a variable to hold the new column name. Add it to the declaration section of the function ODCITableDescribe, on line 12 of the file METHOD4_OT.TPB.

  v_new_column_name varchar2(32);

Create a SQL statement to map the old column to the new column. Add this to line 31, where it will be run for each column.

     --Get mapped column name if it exists.  If none, use the existing name.
     select nvl(max(target_column_name), r_sql.description(i).col_name)
     into v_new_column_name
     from column_names
     where source_column_name = r_sql.description(i).col_name;

Change line 42 to refer to the new variable name:

                      substr(v_new_column_name, 1, 30),

Mapping Table

drop table column_names;
create table column_names
(
    source_column_name varchar2(30),
    target_column_name varchar2(30),
    constraint column_names_pk primary key(source_column_name)
);
insert into column_names values('A1234', 'BETTER_COLUMN_NAME');
insert into column_names values('B4321', 'Case sensitive column name.');

Query Example

Now the column names from any query can magically change to whatever values you want. And this doesn't simply use text replacement; the columns from a * will also change.

SQL> select * from table(method4.query('select 1 a1234, 2 b4321, 3 c from dual'));

BETTER_COLUMN_NAME Case sensitive column name.          C
------------------ --------------------------- ----------
                 1                           2          3

Warnings

Oracle SQL is horrendously complicated and any attempt to build a layer on top if it has many potential problems. For example, performance will certainly be slower. Although I've created many unit tests I'm sure there are some weird data types that won't work correctly. (But if you find any, please create a Github issue so I can fix it.)

In my experience, when people ask for this type of dynamic behavior, it's usually not worth the cost. Sometimes a little extra typing is the best solution.