Query from variable column name?

Running on LAMP
Table - CONTACTS (35 COLUMNS APPROX)
I want to be able to create a results page from a detailed
query.
My first attempt was to include 35 text fields in a form and
then build the query: like '%namesearch%' and like
'%companysearch%' and like '%companytype%' etc. etc. etc.
But this is time consuming and prone to bugs.
I though a much easier way would be to have two fields: one
criteria field and one field representing the column you want to
search in....
Hmm...
This is the query that I thought might work.
SELECT * FROM CONTACTS
WHERE 'colsearch' LIKE '%critsearch%'
ORDER BY COMPANY, FULLNAME
Obviously 'colsearch' is a variable from a menu drop down
field.
Strangely, if I set the default value for colsearch as
COMPANY and default value for critsearch as '%', it does seem to
return a result (all records) on the page load (I have the form and
results on the same page) but if I hit submit, it returns no
recordset!
Is this even possible?
CODE:
$colsearch_recordlookup = "COMPANY";
if (isset($HTTP_POST_VARS['scolumn'])) {
$colsearch_recordlookup = (get_magic_quotes_gpc()) ?
$HTTP_POST_VARS['scolumn'] :
addslashes($HTTP_POST_VARS['scolumn']);
$critsearch_recordlookup = "%";
if (isset($HTTP_POST_VARS['scriteria'])) {
$critsearch_recordlookup = (get_magic_quotes_gpc()) ?
$HTTP_POST_VARS['scriteria'] :
addslashes($HTTP_POST_VARS['scriteria']);
mysql_select_db($database_dbcrm1, $dbcrm1);
$query_recordlookup = sprintf("SELECT * FROM CONTACTS WHERE
'%s' LIKE '%%%s%%' ORDER BY COMPANY, FULLNAME",
$colsearch_recordlookup,$critsearch_recordlookup);

RichardODreamweaver wrote:
> This is the query that I thought might work.
>
> SELECT * FROM CONTACTS
> WHERE 'colsearch' LIKE '%critsearch%'
> ORDER BY COMPANY, FULLNAME
>
> Obviously 'colsearch' is a variable from a menu drop
down field.
The name of the column should not be in quotes:
SELECT * FROM CONTACTS
WHERE colsearch LIKE '%critsearch%'
ORDER BY COMPANY, FULLNAME
David Powers, Adobe Community Expert
Author, "Foundation PHP for Dreamweaver 8" (friends of ED)
Author, "PHP Solutions" (friends of ED)
http://foundationphp.com/

Similar Messages

Result of an SQL query as a Column name of another query

Hi Friends,
Can we use a result of a SQL Query as a column name of another table to retrieve data. If so please help me.
For eg :
I have a table where is store numbers;
select col1 from table1 where col1='5';
and i have another table where .. this value of col is a column name..
select ( select col1 from table1 where col1='5') from table2;
Thanks in advance.

Hi,
ORAFLEX wrote:
Hi Friends,
Can we use a result of a SQL Query as a column name of another table to retrieve data. If so please help me.
For eg :
I have a table where is store numbers;
select col1 from table1 where col1='5';
and i have another table where .. this value of col is a column name..
select ( select col1 from table1 where col1='5') from table2;
Thanks in advance.Do you really mean that?
select col1 from table1 where col1='5';That query will always return either '5' or nothing. Whatever you're trying to accomplish with that, you can do with an EXISTS query.
Perhaps you meant to reference two different columns in that query:
select col1 from table1 where col2='5';In that case, sorry, no, you can't do that without resorting to dynamic SQL.
If the same column is used throughout the query (but could change every time you run the query), then the dynamic SQL might be pretty easy. In SQL*Plus, for example, you could use substitution variables, defined in another query at run-time.
If there are only a few possible values that the sub-query could possibly return, and you know what they all are, then you can fake a dynamic query like this:
SELECT     CASE     ( SELECT col1
            FROM       table1
            WHERE       col2     = '5'
          WHEN 'BONUS'     THEN bonus
          WHEN 'COMM'     THEN comm
          WHEN 'SAL'     THEN sal
     END     AS col1
FROM     table2
;Sorry to give such a vague answer, but it's the best I can do with the information I have.
It would help if you posted a little sample data (CREATE TABLE and INSERT statments for both tables), and the results you want to get from that data. If you want to pass a parameter to the query, give the results you want for a couple of different parameters.

How to make dynamic query using DBMS_SQL variable column names

First of all i will show a working example of what i intend to do with "EXECUTE IMMEDIATE":
(EXECUTE IMMEDIATE has 32654 Bytes limit, which isn't enough for me so i'm exploring other methods such as DBMS_SQL)
-------------------------------------------------CODE-----------------------------------
create or replace PROCEDURE get_dinamic_query_content
(query_sql IN VARCHAR2, --any valid sql query ('SELECT name, age FROM table')
list_fields IN VARCHAR2) --list of the columns name belonging to the query ( arr_list(1):='name';   arr_list(2):='age';
-- FOR k IN 1..arr_list.count LOOP
-- list_fields := list_fields || '||content.'||arr_list(k)||'||'||'''~cs~'''; )
AS
sql_stmt varchar (30000);
BEGIN
               sql_stmt :=
'DECLARE
     counter NUMBER:=0;
auxcontent VARCHAR2(30000);
     CURSOR content_cursor IS '|| query_sql ||';
     content content_cursor%rowtype;
     Begin
          open content_cursor;
          loop
               fetch content_cursor into content;
               exit when content_cursor%notfound;
               begin
                    auxcontent := auxcontent || '||list_fields||';
               end;
               counter:=counter+1;
          end loop;
          close content_cursor;
          htp.prn(auxcontent);
     END;';
EXECUTE IMMEDIATE sql_stmt;
END;
-------------------------------------------------CODE-----------------------------------
I'm attepting to use DBMS_SQL to perform similar instructions.
Is it possible?

Hi Pedro
You need to use DBMS_SQL here because you don't know how many columns your query is going to have before runtime. There are functions in DBMS_SQL to get information about the columns in your query - all this does is get the name.
SQL> CREATE OR REPLACE PROCEDURE get_query_cols(query_in IN VARCHAR2) AS
2 cur PLS_INTEGER;
3 numcols NUMBER;
4 col_desc_table dbms_sql.desc_tab;
5 BEGIN
6 cur := dbms_sql.open_cursor;
7 dbms_sql.parse(cur
8 ,query_in
9 ,dbms_sql.native);
10 dbms_sql.describe_columns(cur
11 ,numcols
12 ,col_desc_table);
13 FOR ix IN col_desc_table.FIRST .. col_desc_table.LAST LOOP
14 dbms_output.put_line('Column ' || ix || ' is ' ||
15 col_desc_table(ix).col_name);
16 END LOOP;
17 dbms_sql.close_cursor(cur);
18 END;
19 /
Procedure created.
SQL> exec get_query_cols('SELECT * FROM DUAL');
Column 1 is DUMMY
PL/SQL procedure successfully completed.
SQL> exec get_query_cols('SELECT table_name, num_rows FROM user_tables');
Column 1 is TABLE_NAME
Column 2 is NUM_ROWS
PL/SQL procedure successfully completed.
SQL> exec get_query_cols('SELECT column_name, data_type, low_value, high_value FROM user_tab_cols');
Column 1 is COLUMN_NAME
Column 2 is DATA_TYPE
Column 3 is LOW_VALUE
Column 4 is HIGH_VALUE
PL/SQL procedure successfully completed.I've just written this as a procedure that prints out the column names using dbms_output - I guess you're going to do something different with the result - maybe returning a collection, which you'll then parse through in Apex and print the output on the screen - this is just to illustrate the use of dbms_sql.
best regards
Andrew
UK

Variable column name in a trigger

I have a table with many columns, and would like to make a log table of all the updates in the first one. So my idea was to bulid a "before update" trigger which would first create a cursor to get the column names from the user_tab_columns table. In a cursor loop it would check if the :new.column_data is somehow different from :old.column_data. If this was the case, trigger would build a sql statement for inserting old and new values into the log table, and then execute this statement using the "execute immediate".
Here i run into a problem, because i shuld pass the old and new values of a column instead of :new and :old variables, which are not bound in the execute immediate. Any suggestion besides writing a long trigger for checking each and every column name in a table?

Or don't create it like this at all.
This way you'll get:
- complex trigger code
- a hard parse each and every insert into that log table
- a trashed shared pool, because it gets populated with all "almost the same" insert statements
Querying a log table probably is not something you do everyday, so it likely doesn't pay to store it in a complex way to make sure the retrieval goes smoothly.
If you want to log, you could do it in a single insert statement with all the information you could possibly need. This makes the code easy and the SQL sharable. It could look like below, and just skip the columns you don't need:
insert into log_table
( action
, action_date
, action_user
, column1_old_value
, column1_new_value
, did_column1_change
, column2_old_value
, column2_new_value
, did_column2_change
, columnN_old_value
, columnN_new_value
, did_columnN_change
values
( case when inserting then 'I' when updating then 'U' when deleting then 'D' end
, sysdate
, user
, :old.column1
, :new.column1
, case when updating('column1') then 'Y' else 'N' end
, :old.column2
, :new.column2
, case when updating('column1') then 'Y' else 'N' end
, :old.columnN
, :new.columnN
, case when updating('columnN') then 'Y' else 'N' end
);Or you can use a flashback archive instead of a log table in 11g, but I don't have enough experience with it yet to fully recommend that one.
Regards,
Rob.

Order query depending on column names passed at runtime

Hi
Can any one suggest how to order by at run time depending on column name passed as parameter.
create or replace procedure dummy_proc( p_orderby varchar2,p_refcursor out sys_refcursor)
is
begin
     open p_refcursor for select * from
     emp where Status='A'
          order by (select p_orderby from dual);
end;
--execution part for testing
declare
     v_outcursor sys_refcursor;
     v_rec_data emp%rowtype;
     v_cnt number;
begin
     dummy_proc ('ename,sal',v_outcursor);
     begin
          loop
               fetch v_outcursor into v_rec_data;
               v_cnt:=v_outcursor%rowcount;
               exit when v_outcursor%notfound;
               dbms_output.put_line('empno '|| to_char(v_rec_data.empno));
               dbms_output.put_line('empno '|| v_rec_data.ename);
          end loop;
     end;
     dbms_output.put_line('Count is'|| to_char(v_cnt));
end;
we do not want to something like below as my actual query is very big with union statements and case expressions, (below is just an similar example)
how i tried to order by but its not working.
v_sql='select * from emp where empno= '|| p_empno || ' order by '|| p_orderby;
open p_refcursor for v_sql;
Thanks in advance
Vishal

we do not want to something like below as my actual
query is very big with union statements and case
expressions, (below is just an similar example)
how i tried to order by but its not working.
v_sql='select * from emp where empno= '|| p_empno ||
' order by '|| p_orderby;
open p_refcursor for v_sql;I agree, you really don't want to do that, you want to use BIND VARIABLES.
v_sql := ' select * from emp where empno = :bound_empno order by :bound_orderby';
open p_refcursor for v_sql using p_empno, p_orderby;
[pre]
The size of the statement shouldn't come in to play, if the only dynamic portion of it is the order by clause then declare a constant which is the large query (you can use a LONG datatype) and then concatenate in the order by clause before you return the reference cursor.
Unless you can give a legitimate reason for why you cannot use Dynamic SQL (not why you do not want to use it) this would be my approach.
And please, if you don't know what a bind variable is...read about them, or your system will suffer.

Variable column names

I'm working on a data collection application. Actual device names to be logged can change. Two tables are used: a log "header" table that contains an entry with input descriptions (one record per "run") and another table that contains logged data every 10 secondss
Log header table:
DT | run_num | input1_desc | input2_desc |
xxxx 1234 inlet_temp zone_1_temp
Data table:
DT | run_num | input1 | input2 |
xxxx 1234 12.234 45.675
How can I create a view to show data with actual column names (ex: inlet_temp)? Or should db be re-designed?

Are you using an application express report to display the results?
If so, you could use the custom pl/sql headings for that purpose.
This example could get you started:
DECLARE
   l_ret   VARCHAR2 (500);
BEGIN
   SELECT 'DT:RUN_NUM:' || input1_desc || ':' || input2_desc
     INTO l_ret
     FROM logheader
    WHERE run_num = :p1_run_num;
   RETURN l_ret;
END;These custom headings only make sense for a single run_num.
So create a report for the details :
select * from data where run_num = :p1_run_num; and use the above mentioned pl/sql block on the report attributes tab for a heading type of pl/sql.
The variable :p1_run_num refers to a variable on page 1 on which you set the current run_num value for the details.
Regards,
~Dietmar.

How to get the alias name of an attribute of a VO from its column name?

How can I programmatically get the alias name of an attribute of a VO given its column name?
Often alias and column names are not same when the VO is created (and I don't require them to be the same)

Hello John.
I will be more precise. When the VO based on a table comes out from the wizard, it produces xml entries like the following one:
<ViewAttribute
Name="ColumnName"
AliasName="COLUMN_NAME"/>There the column of the table on the DB really is COLUMN_NAME. The wizard mangles that name producing ColumnName as Name property.
This is ok. I don't want to change this default behaviour.
What I need is to programmatically get the value of that attribute given its AliasName property value:
String f(ViewObject vo, String AliasName) {
return (String)vo.getCurrentRow().getAttributeFromAliasName(aliasName); // getAttributeFromAliasName does not exist
// instead of return (String)vo.getCurrentRow().getAttribute(Name);
That is, usually getAttribute expects what in the xml is the Name property. However, I want a sort of getAttribute() expecting the AliasName.
I hope it's clear.

Create App From Spreadsheet: "column name is an Oracle reserved word"

I'm creating an App from a spreadsheet on apex.oracle.com (3.0.0.00.20 at time of writing) and receiving the error "The identified column name is an Oracle reserved word. Please choose another name."
I've tried adding an underscore to the end of each column name in order to avoid having to check every column name (yup, lazy) but that didn't remove the error, nor is any column name identified by the interface.
Thoughts?

Hi David,
There was a bug at one time where reserved words were not being checked properly when creating columns. A particular problem related to names beginning with "SYS".
Try adding the underscore in front of the column name.
Regards
Andy

Query to return column name of first NULL column?

I have a table with 40 columns but only the first n have non-null values. Is there a way to pull the name of the first non-null column? I have tried using CASE but I run into nesting too deeply problems (apparently you can only nest to 10 levels).
Thank you.
Kevin
Kevin Burton

Or, you could try something like this:
-- This is just for testing
DROP TABLE #test
CREATE TABLE #Test
(ID INTEGER
,n1 INTEGER
,n2 INTEGER
,n3 INTEGER
,n4 INTEGER
,n5 INTEGER
,n6 INTEGER
,n7 INTEGER
,n8 INTEGER
,n9 INTEGER
INSERT INTO #Test
VALUES (1,256,365,4000,0,NULL,NULL,NULL,NULL,NULL)
SELECT *
FROM #Test
-- THIS IS THE IMPORTANT PIECE
SELECT ISNULL(LEN(LEFT(n1,1)),0)
+ ISNULL(LEN(LEFT(n2,1)),0)
+ ISNULL(LEN(LEFT(n3,1)),0)
+ ISNULL(LEN(LEFT(n4,1)),0)
+ ISNULL(LEN(LEFT(n5,1)),0)
+ ISNULL(LEN(LEFT(n6,1)),0)
+ ISNULL(LEN(LEFT(n7,1)),0)
+ ISNULL(LEN(LEFT(n8,1)),0)
+ ISNULL(LEN(LEFT(n9,1)),0)
FROM #Test
The LEFT(x,1) will always either give you a LEN of 1, or a NULL. Add 'em up. Try it.
Then use the Ordinal_Position from the Information Schema to get the name of the column (if that is what you need)
Duncan Davenport

Unpivot taking the last letter from the column name

Hi guys, having
DATE               Period1 Period2 Period3 Period4
01/01/2014       1.02       1.06      1.02      1.03
02/01/2014       1.06       1.05      1.04      1.06
I would like an unptivot as
01/01/2014    1        1.02
01/01/2014    2        1.06
01/01/2014    3        1.02
01/01/2014    4        1.03
02/01/2014    1        1.06
02/01/2014    2        1.05
02/01/2014    3        1.04
02/01/2014    4        1.06
But I'm wondering how can I take the last letter of the column. Maybe I should use a select case...
Any advices? Thank you

DECLARE @Table TABLE
dt DATE,
P1 INT NULL,
P2 INT NULL,
P3 INT NULL,
P4 INt NULL
INSERT INTO @Table(dt, P1, P2, P3, P4)
VALUES('20140101',10,null,20,3);
INSERT INTO @Table(dt, P1, P2, P3, P4)
VALUES('20140102',50,25,15,5);
SELECT dt, row_number() over (partition by dt order by dt) rn,per AS DayMax FROM @Table
UNPIVOT (per FOR DayNumber IN (P1, P2, P3, P4)) AS c
Best Regards,Uri Dimant SQL Server MVP,
http://sqlblog.com/blogs/uri_dimant/
MS SQL optimization: MS SQL Development and Optimization
MS SQL Consulting:
Large scale of database and data cleansing
Remote DBA Services:
Improves MS SQL Database Performance
SQL Server Integration Services:
Business Intelligence

Data load from variable file names

I have multiple files that I want to load into a cube, each starting with the same 5 characters but ending differently. EG GM1010104 GM1010204 What's the best option for MAXL script to automate this data load ? Can you use a wildcard name in the script to pick up anything starting with GM101**** ?

No - you need to specify the file name as it appears properly (I've never tried it but I am pretty sure it wouldn't work). One solution to this problem though is to have a shell script (or DOS commands) auto-generate an ESSCMD/MaxL script based on the files that exist in a directory.Most scripting environments should allow you to loop through a list of files that match some pattern - you can then create a script with the results and execute it.Another option is to build a MaxL script that accepts a parameter (file name) and have a shell script call it as it loops through the file list.Hope that helps.Regards,Jade----------------------------------Jade ColeSenior BI ConsultantClarity [email protected]

Passing column names in as a variable

I know in SQL server there is a way you can pass a variable in to the procedure and put that variable as a column name, Oracle doesn't seem to pick the variable up as a column name. Can anyone show me how with a simple select statement

bgulcu wrote:
I'm having a similar problem. I didn't want to open up another thread since it's related.
Problem:
I have 107 columns on a table. The number of columns increases almost every month.
I have an array of columns, which I have gathered from the db. I need to pivot the whole table, but due to indefinite number of columns, it had to be dynamic sql.
Premature Solution:
Right now, the solution that we have runs through the array of columns and uses dynamic sql to query for every single cell to get the value. Considering that we have 1000+ rows on this table, in order to do the transformation, we send 107000+ queries which is just ugly. It works, but I don't feel comfortable about it.
Solution:
Since we are able to get a single row of 107 columns, why don't we gather the whole row, and get the information out from it instead of tiring up the server? The indefinite number of columns is a problem, but we have the whole list in an array, lets say feature_array. So what I want to do is to get the date from the row, ie myRow.feature_array(1), myRow.feature_array(2), ..., myRow.feature_array(i).
I just learned that it's not possible to do it (from this thread).You just misunderstood the contents of this thread. What we said is that you can't do such things with static SQL, so you need to use dynamic sql.
What has also been told was that using dynamic sql for such a purpose is just the result of a bad design. Use a relational table as a bi-dimensional array is not right, it is possible but it's not right for the simple reason that you need a procedural language to process the dynamic SQL that is needed to retrieve data.
Probably on database programming there's only a rule of thumb and it says: "Maximize SQL minimize PL/SQL" (or any procedural language like java). The only exceptions to keep code in procedural language must depend on a complexity reduction determined by its use. In your case there is no complexity reduction determined by the use of a procedural language.
In the thread on this link(Schema integrity I discussed about how to implement a solution you can use as alternative for submitting queries with variable column names.
If you introduce parameter management in your model probably you could have your problem solved with the use of only a single SQL statement. And that may be a good designed model.
Bye Alessandro

Querying MSAccess with # in column name

What is the proper syntax for a JDBC sql statement to query a MSAccess database with "#' (the pound or number sign) in a column name?
Thanks.
-John

Thanks all for the speedy replies. This forum should be extremely helpful in the future.
The double quotes did the job. I thought I had tried using quotes, but I must have used single quotes instead.
I do not have a say in the database design, and therefore, am stuck using whatever is thrown at me. Otherwise, I agree that staying away from using column names that may not conform to accepted standards is a good idea.
Thanks again.
-John

Using variable coulmn name in sql function

Hi there,
I am not an expert with PL/SQL and I can not figure out how to use variable column names in my function.
My function is:
CREATE OR REPLACE FUNCTION RESET_TRIGGERS(aTrigger VARCHAR2) RETURN NUMBER IS
TEMP_ID NUMBER;
TEMP_USER_ID NUMBER;
BEGIN
SELECT 'LIMS.'||'$aTrigger'||'.NEXTVAL' INTO TEMP_ID FROM DUAL;
SELECT 'LIMS.'||'$aTrigger'||'_USER.NEXTVAL' INTO TEMP_USER_ID FROM DUAL;
IF TEMP_ID > TEMP_USER_ID THEN
LOOP
SELECT LIMS.SQ_U_FINALRESULT_USER.NEXTVAL INTO TEMP_USER_ID FROM DUAL;
EXIT WHEN TEMP_USER_ID = TEMP_ID;
END LOOP;
ELSE
WHILE TEMP_ID < TEMP_USER_ID LOOP
SELECT LIMS.SQ_U_FINALRESULT.NEXTVAL INTO TEMP_ID FROM DUAL;
END LOOP;
END IF;
COMMIT;
RETURN (TEMP_ID);
END;
What I want is that I pass a seqencename with aTrigger and that two triggers will be equal if not.
eg ifaTrigger = 'SQ_U_FINALRESULT'
than I want the triggers LIMS.SQ_U_FINALRESULT and LIMS.SQ_U_FINALRESULT_USER to be set equal.
The above function will not work, but what will?????
I hope you can help me out!
Cheers

A very strange function indeed.
But here is what I think he meant to do:
SQL> create procedure reset_sequences
2 ( p_sequence_name in varchar2
3 , p_nextval          out number
4 )
5 is
6    l_nextval1 number;
7    l_nextval2 number
8    ;
9    procedure reset_sequence_value
10    ( p_sequence_name in varchar2
11    , p_current_value in number
12    , p_new_value     in number
13    )
14    is
15      l_dummy number;
16    begin
17      execute immediate 'alter sequence ' || p_sequence_name || ' increment by ' || to_char(p_new_value-p_current_value);
18      execute immediate 'select ' || p_sequence_name || '.nextval from dual' into l_dummy;
19      execute immediate 'alter sequence ' || p_sequence_name || ' increment by 1';
20    end reset_sequence_value
21    ;
22 begin
23    execute immediate
24      'select ' || p_sequence_name || '.nextval,' || p_sequence_name || '_user.nextval from dual'
25    into l_nextval1, l_nextval2
26    ;
27    if l_nextval1 < l_nextval2
28    then
29      reset_sequence_value(p_sequence_name,l_nextval1,l_nextval2);
30    end if
31    ;
32    if l_nextval1 > l_nextval2
33    then
34      reset_sequence_value(p_sequence_name || '_user',l_nextval2,l_nextval1);
35    end if
36    ;
37    p_nextval := greatest(l_nextval1,l_nextval2)
38    ;
39 end reset_sequences;
40 /
Procedure is aangemaakt.
SQL> show err
Er zijn geen fouten.
SQL> create sequence testseq start with 5 increment by 1
2 /
Reeks is aangemaakt.
SQL> create sequence testseq_user start with 2 increment by 1
2 /
Reeks is aangemaakt.
SQL> declare
2    l_new_value number;
3 begin
4    reset_sequences('testseq',l_new_value);
5    dbms_output.put_line(l_new_value);
6 end;
7 /
5
PL/SQL-procedure is geslaagd.
SQL> select testseq.currval from dual
2 /
                               CURRVAL
                                     5
1 rij is geselecteerd.
SQL> select testseq_user.currval from dual
2 /
                               CURRVAL
                                     5
1 rij is geselecteerd.Regards,
Rob.

Invalid Column Name message in the subquery

Hi All, I am receiving the invalid column name message in the subquery. Please help me if there is any work around or change in the query:
Sample query:
update table1 set table1.column1 =
(Select column1 from
(Select column1 from table2 where
table1.columnX = table2.columnX
order by table2.columnM desc,table2.ColumnN desc,
where rownum =1)
When i run the above query i am receiving the following error message:
ORA-00904:invalid column name at table1.columnx.
This column is already existing in table1 and looks like this table is not recoginzed inside of the subquery.
I am trying to update table1.column1 using the first value from table2 after ordering by table2.columnM desc, table2.columnN desc. (ColumnM and ColumnN are the Date columns which i need to perform the sorting). I am new to sql/Plsql. Thanks in Advance for help....

The problem is that you can not refer to columns from the outer table from inside a second subquery inside your subquery. Only one level of subquery can be used to transport the column name,.
A solution can be to use groups and aggregation functions instead of rownum = 1.
The following might work.
untested
update table1
set table1.column1 =
(Select max(column1) keep (dense_rank first over table2.columnM desc,table2.ColumnN desc)
   from table2
   where table1.columnX = table2.columnX
   ) Another way can be to to make a UNCORRELATED subquery first and then filter on that uncorrelated query with the column name from the outer statement.
This time I use a analytic function.
untested
update table1 u
set u.column1 = (select v1.column1
   from
     (Select t2.column1, t2.columnX, row_number over (partition by t2.column1, t2.columnX order by t2.columnM desc,t2.ColumnN desc) as rn
      from table2
      ) v1
   where v1.columnX = u.columnX
   and v1.rn = 1
where /* where clause is missing! */

Query from variable column name?

Similar Messages

Maybe you are looking for