在查询标头中设置DATE Oracle SQL（SQL Developer）

Question

I am trying to display specific data that will be using DATE in headers.

Data should be displayed in similar format to the shown below.

Master_Column|Todays_date|Today-1|Today-2|Today-3...

select  
   sysdate,
   SYSDATE "&&_DATE",
   (sysdate-1) "&&_DATE-1" --this header is incorrectly named
from 
  dual;

This however will not display the last header correctly.

Is there any way to set dates in headers that will change dynamically?

I need to have a range of 6 days. Each day as separate column.

Answer 1

The _DATE substituion variable is a string, which tracks your NLS_DATE_FORMAT:

show defines

DEFINE _DATE =  "2018-10-31" (CHAR)

You're sort of trying to subtract a number from a string; except you aren't even really doing that, because no calculation would be done inside the double-quotes anyway.

It's a bit messy but you could achieve the effect I think you are after using multiple substituion variables and the column ... new_value ... syntax supported by SQL Developer, with a query to set those values - which you can hide with set termout off (at least in recent versions):

column date_0 new_value date_0
column date_1 new_value date_1
column date_2 new_value date_2
-- ...

set termout off
select to_char(sysdate, 'YYYY-MM-DD') as date_0,
  to_char(sysdate - 1, 'YYYY-MM-DD') as date_1
  -- ...
from dual;
set termout on

set verify off

select  
   sysdate,
   SYSDATE "&date_0",
   (sysdate-1) "&date_1"
from 
  dual;

SYSDATE    2018-10-31 2018-10-30
---------- ---------- ----------
2018-10-31 2018-10-31 2018-10-30

However, you need to put the whole thing, or at least the 'hidden' part, into a file and then run that with @file.sql , as termout ins't honoured for statements run directly from the worksheet. ALternatively you could add noprint to the column definitions, which hides them too - but give syou extra blank lines in the output still.

You can also use those substituion variables to change the column formatting if you need to, eg with shorter dummy values selected:

column master_column format a20
column &date_0 format A10
column &date_1 format A15

select  
   'a' as master_column,
   'b' as "&date_0",
   'c' as "&date_1"
from 
  dual;

MASTER_COLUMN        2018-10-31 2018-10-30     
-------------------- ---------- ---------------
a                    b          c              

Note that in the column &date_0 ... etc. clauses the variable names should not be enclosed in double quotes, unlike the aliases in the main query which do have to be quoted (because of the characters used in and the formatting of the alias name).

You can extend that approach further, and perhaps more neatly, by setting the column headers that way too; so the query itself doesn't have to worry about them:

column master_column format a20
column col2 format A10 heading &date_0
column col3 format A15 heading &date_1

select  
   'a' as master_column,
   'b' as col2,
   'c' as col3
from 
  dual;

MASTER_COLUMN        2018-10-31 2018-10-30     
-------------------- ---------- ---------------
a                    b          c