Selecting fewer columns from cached results

If I'm reading the docs right, I should be able to create a request with X number of columns to populate the cache, so that a subsequent request that has a direct subset of those columns would hit the cache. That's not what I'm seeing.
I have four fact columns in the RPD. The first three, Sales, Cost, and Units, are coming straight from the Physical Layer. The fourth, Profit, is a subtraction of Sales minus Cost.
Here's what I'm seeing:
1a. Log In to Answers
1b. Create Query: Year, Sales, Cost, Profit - this populates the cache.
1c. Log Out of Answers
1d. Log back in to Answers
1e. Create Query: Year, Sales, Cost - this query does NOT hit that cache entry.
2a. Close all cursors (under Administration...Manage Sessions)
2b. Log Out
2c. Purge Cache
3a. Log In to Answers
3b. Create Query: Year, Sales, Cost, Units - this populates the cache.
3c. Log Out of Answers
3d. Log back in to Answers
3e Create Query: Year, Sales Cost - this query DOES hit the cache entry.
The only difference between the two sequences is that one references a calculated column (Profit) and the other does not.
Incidentally, I've tested this with Profit defined first as a subtraction of the logical columns, and then again as a subtraction of the physical columns, with the same behavior observed both ways.
Can someone please confirm this behavior and suggest a reason why it should occur?

Turribeach, here are the Session Logs for each of the four requests. I've interspersed the steps between double-rows of = signs.
===================================================
===================================================
STEP 1A: Log In
STEP 1B: Create request - Year, Sales, Cost, Profit
===================================================
===================================================
+++Analyst:320000:320001:----2009/01/15 09:13:51
-------------------- Logical Request (before navigation):
RqList
Times.Year as c1 GB,
Sales:[DAggr(~Base Facts.Sales by [ Times.Year, Times.Year End Date] )] as c2 GB,
Cost:[DAggr(~Base Facts.Cost by [ Times.Year, Times.Year End Date] )] as c3 GB,
Sales:[DAggr(~Base Facts.Sales by [ Times.Year, Times.Year End Date] )] - Cost:[DAggr(~Base Facts.Cost by [ Times.Year, Times.Year End Date] )] as c4 GB,
Times.Year End Date as c5 GB
OrderBy: c5 asc
+++Analyst:320000:320001:----2009/01/15 09:13:51
-------------------- Sending query to database named BIEE_TRAIN (id: <<1690>>):
WITH
SAWITH0 AS (select sum(round(T68.COST , 2)) as c1,
sum(round(T68.SALES , 2)) as c2,
T61.YEAR as c3,
T61.YEAR_END_DATE as c4
from
GLOBAL_ADMIN.BI_D_TIME T61,
GLOBAL_ADMIN.BI_F_SALES T68
where ( T61.MONTH = T68.MONTH )
group by T61.YEAR, T61.YEAR_END_DATE)
select distinct SAWITH0.c3 as c1,
SAWITH0.c2 as c2,
SAWITH0.c1 as c3,
SAWITH0.c2 - SAWITH0.c1 as c4,
SAWITH0.c4 as c5
from
SAWITH0
order by c5
+++Analyst:320000:320001:----2009/01/15 09:13:52
-------------------- Query Result Cache: [59124] The query for user 'Analyst' was inserted into the query result cache. The filename is 'C:\OracleBIData\cache\NQS_TRAINING_733424_33231_00000000.TBL'.
===================================================
===================================================
STEP 1C: LOG OUT
STEP 1D: LOG IN
STEP 1E: Create request - Year, Sales, Cost
===================================================
===================================================
+++Analyst:330000:330001:----2009/01/15 09:17:41
-------------------- Logical Request (before navigation):
RqList
Times.Year as c1 GB,
Sales:[DAggr(~Base Facts.Sales by [ Times.Year, Times.Year End Date] )] as c2 GB,
Cost:[DAggr(~Base Facts.Cost by [ Times.Year, Times.Year End Date] )] as c3 GB,
Times.Year End Date as c4 GB
OrderBy: c4 asc
+++Analyst:330000:330001:----2009/01/15 09:17:41
-------------------- Sending query to database named BIEE_TRAIN (id: <<2062>>):
select T61.YEAR as c1,
sum(round(T68.SALES , 2)) as c2,
sum(round(T68.COST , 2)) as c3,
T61.YEAR_END_DATE as c4
from
GLOBAL_ADMIN.BI_D_TIME T61,
GLOBAL_ADMIN.BI_F_SALES T68
where ( T61.MONTH = T68.MONTH )
group by T61.YEAR, T61.YEAR_END_DATE
order by c4
+++Analyst:330000:330001:----2009/01/15 09:17:41
-------------------- Query Result Cache: [59124] The query for user 'Analyst' was inserted into the query result cache. The filename is 'C:\OracleBIData\cache\NQS_TRAINING_733424_33461_00000001.TBL'.
*** Note: At this point, there are indeed TWO entries in the cache. ***
===================================================
===================================================
STEP 2A: Close all cursors
STEP 2B: Log Out
STEP 2C: Purge the cache
STEP 3B: Create request - Year, Sales, Cost, Units
===================================================
===================================================
+++Analyst:350000:350001:----2009/01/15 09:23:18
-------------------- Logical Request (before navigation):
RqList
Times.Year as c1 GB,
Sales:[DAggr(~Base Facts.Sales by [ Times.Year, Times.Year End Date] )] as c2 GB,
Cost:[DAggr(~Base Facts.Cost by [ Times.Year, Times.Year End Date] )] as c3 GB,
Units:[DAggr(~Base Facts.Units by [ Times.Year, Times.Year End Date] )] as c4 GB,
Times.Year End Date as c5 GB
OrderBy: c5 asc
+++Analyst:350000:350001:----2009/01/15 09:23:18
-------------------- Sending query to database named BIEE_TRAIN (id: <<2399>>):
select T61.YEAR as c1,
sum(round(T68.SALES , 2)) as c2,
sum(round(T68.COST , 2)) as c3,
sum(round(T68.UNITS , 0)) as c4,
T61.YEAR_END_DATE as c5
from
GLOBAL_ADMIN.BI_D_TIME T61,
GLOBAL_ADMIN.BI_F_SALES T68
where ( T61.MONTH = T68.MONTH )
group by T61.YEAR, T61.YEAR_END_DATE
order by c5
+++Analyst:350000:350001:----2009/01/15 09:23:19
-------------------- Query Result Cache: [59124] The query for user 'Analyst' was inserted into the query result cache. The filename is 'C:\OracleBIData\cache\NQS_TRAINING_733424_33798_00000002.TBL'.
===================================================
===================================================
STEP 3C: LOG OUT
STEP 3D: LOG IN
STEP 3E: Create request - Year, Sales, Cost
===================================================
===================================================
+++Analyst:360000:360001:----2009/01/15 09:24:36
-------------------- Logical Request (before navigation):
RqList
Times.Year as c1 GB,
Sales:[DAggr(~Base Facts.Sales by [ Times.Year, Times.Year End Date] )] as c2 GB,
Cost:[DAggr(~Base Facts.Cost by [ Times.Year, Times.Year End Date] )] as c3 GB,
Times.Year End Date as c4 GB
OrderBy: c4 asc
+++Analyst:360000:360001:----2009/01/15 09:24:36
-------------------- Cache Hit on query:
Matching Query: SET VARIABLE QUERY_SRC_CD='Report';SELECT "Time Dimension"."Year" saw_0, "Base Measures".Sales saw_1, "Base Measures".Cost saw_2, "Base Measures".Units saw_3 FROM Global ORDER BY saw_0
Created by: Analyst
+++Analyst:360000:360001:----2009/01/15 09:24:36
-------------------- Query Status: Successful Completion
+++Analyst:360000:360001:----2009/01/15 09:24:36
*** Note: At this point, there is ONE entry in the cache. ***

Similar Messages

SQL*Loader - How to load only a few columns from a .csv file to ora table

Hi there,
Could anyone please let me know how to load few columns from a .csv
file into a oracle table using SQL*Loader.
I know how to create a .dat and .ctl file and run the sql loader.
Suppose I have a .csv file with
col1, col2, col3, col4
and I only need to load col1 and col3 into col_a and col_b respectively
in table_a?
structure of table_ a
col_a,
col_b
Please advice

Try like..it i will work..
LOAD DATA
INFILE 'test.txt'
LOAD DATA
TRUNCATE INTO TABLE T1
FIELDS TERMINATED BY ','
(col1,
col2 FILLER,
col3,
col4 FILLER
)

How to Remove totals on few columns from SEM BPS layout

On SEM BPS layout it has 10 5 charatestics columns
and 10 key figure columns.
out of these key figures I want see totals on some
coulmns and for few columns I do not want to see totals.
for wxample it does not make sense of seeing total under
the salary incrase percentage coulmn.
Please suggest anyway we can remove total for some
columns on the layout

Hey, thanks -- everything worked as you described. It never would have occurred to me that the "put back" button refers to the browser selection not the layout selection, for reasons I give below.
I think Aperture exhibits some conceptual confusion here. First of all, a "light table" is created by the New > Light Table command, and you add images to the newly created light table just as you would add them to an album. But those images don't go on the light table layout until you drag them from the browser to the layout table. The "put back" button and "remove from light table" contextual menu command both mean remove from the layout. The phrasing of the contextual menu command make it sounds as if it will remove the clicked image from the light table's collection of images, like removing an image from an album -- it would have made more sense to name this command "put back". Furthermore, the "put back" button is above the layout display, not above the browser, which to me implies that it applies to selections in the layout display. To select an image in the browser, and click "put back selected" seems backwards -- if you just selected the image in the browser, your attention is focused on the browser, not the layout; from the browser's point of view, the button should read "bring back" not "put back".
The documentation didn't really help when I was trying to figure this out. Sometimes the Aperture documentation refers to just a "light table", as on the overview on page 732. On pages 733 and 734, though, it refers to a "light table album". On page 735 of the Aperture manual there is a very brief explanation of how "to add images to the light table" and "to remove an image from the light table". Here, "light table" refers to the layout not the album. The explanation of the "Put Back Selected" button says to select an image then click the button, and the picture shown is of an image selected in the layout, not in the browser.

Selecting all columns from table in a model query

I have written a model query which joins 4 tables, applies some rules and returns updated rows by selecting 4 columns out of this. Currently it works fine because all 4 columns are from MODEL aliases. Now I need to select 16 more columns in the same query but none of these columns is added in MODEL aliases. When I tried selecting these columns I got oracle error ORA - 32614. Can someone please guide me to include these columns in the same model query?
I tried couple of options but no luck. Here are those options for ready reference:
1. I cannot nest existing model query into another select because there are no columns avaiable to join in output of current query to map with other records.
2. I cannot include all 16 columns in MODEL aliases because some of these columns are actually output of user defined functions.
I am using Oracle database version 11g Release 11.2.0.1.0.
Edited by: Anirudha Dhopate on Jan 23, 2011 5:43 PM

Thank you Avijit for your reply. There is a syntax error on the ON in this part of the statement which I don't know how to fix - I tried messing around with another INNER JOIN but am not confident that I'm doing the right thing:
SENAlertType.SENAlertTypeIDONClassMember.ClassMemberStudentID
Thanks for your help! I will need to do some more bedtime reading on joins.
Daniel

Select LONG column from Remote table

Hi to all.
I have the following problem: I can't select a LONG column from remote database, when there are Join operation in the query. Example:
SELECT long_column FROM remote_table@DB
-> that's OK, but:
SELECT long_column FROM remote_table@DB INNER JOIN remote_table2@DB ON ...
-> returns: "ORA-00997: Illegal use of LONG datatype"
I cannot even perform (because there a LONG column in the joined table):
SELECT nonlong_column FROM remote_table@DB INNER JOIN remote_table2@DB ON ...
That's very strange to me because:
SELECT long_column FROM local_table INNER JOIN local_table2 ON ...
-> is OK!
Can someone help me to SELECT a long column from remote in SELECT query with JOIN clause!
Thanks a lot!

Hi
"Distributed queries are currently subject to the restriction that all tables locked by a FOR UPDATE clause and all tables with LONG columns selected by the query must be located on the same database. " by otn docs.
I have no idea.
Ott Karesz
http://www.trendo-kft.hu

Selecting a column from a table conditionally

I'm trying this select statement (example):
select np.phone_number from name_phone np, reservation_general_view rgv
where rgv.resort='HVILLA'
In this table there is a column called "phone_type" and I only want my column to display values that are associated to the column when the "phone_type" = 'EMAIL'
I tried this:
select np.phone_number(select np.phone_number from name_phone np where np.phone_type='EMAIL')
from name_phone np, reservation_general_view rgv
where rgv.resort='HVILLA'
This doesn't seem to work...can someone tell me how to do that?
Message was edited by:
JonnyOpera

select phone_number from name_phone
where resort='HVILLA'
and phone_type='EMAIL'

Select sql running very long when made to select many columns from a table

Hi,
I am using an Oracle DB 10g. I have a table with 250 columns and I need to select 200 columns out of them. The table holds around 0.1 million rows. But when I run the select sql it takes 15 mins to return .1 million rows. Where as if I select only 10 or 15 columns the sql runs in less than a minute returning same number of rows. My sql looks like below:
select p.col1,p.col2,.......,p.col200
from table Parent p;
The table also has a Primary key Index but it does not seem to be using it even when I forced an Index hint. Could you pls help?

961796 wrote:
I am using an Oracle DB 10g. I have a table with 250 columns and I need to select 200 columns out of them. The table holds around 0.1 million rows. But when I run the select sql it takes 15 mins to return .1 million rows. Where as if I select only 10 or 15 columns the sql runs in less than a minute returning same number of rows. My sql looks like below:
As Sven points out, it is likely that most of your time is network (and client) time. You are sending 20 times as much data (based on column counts, at any rate) across the network, and concerned that it's taking 15 times as long.
If you're testing from SQL*Plus then setting the arraysize to a value larger than the default might help.
If you have control over the SQL*Net settings then you may get some benefit by adjusting the SDU sizes at both ends of the link.
If you have control over the tcp configuration (transmit and receive buffer sizes) then you may get some benefit by adjusting these.
Simple test, by the way, if you're on SQL*Plus
set autotrace traceonly statistics
select ...This will dump your data in the bit bucket as it arrives giving you
a) the database time plus network time
b) some statistics including the volume of data down the network and the number of network round-trips that Oracle saw.
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
Author: Oracle Core

Select multiple columns from tab with group by

hi
i have a table with many columns suppose in that i have 2 columns as date and time
and in my table i am getting suppose data like
columns -> date time a b c
22-may-2207 10
22-may-2207 17
23-may-2007 5
24-may-2007 12
24-may-2007 15
i want to select all columns with data with unique data with max of time for that particular date
answer :-
22-may-2007 17 a b c
23-may-2007 5 a b c
24-may-2007 15 a b c
with all other cloumns also with this
plz help me i am trying too much and not able to solve this.

try this
select datecol, timecol, a, b, c
from (
select datecol, timecol, a, b, c,
rank() over(partition by datecol order by timecol desc) rn
from my_table
where rn = 1
order by datecol;Message was edited by:
Jens Petersen

Urgent...... how to select few fields, from database table, (dynamic ita.)

Dear all experts,
I am able to populate all fields data from database table, using dynamic table creation.
eg,
SELECT * FROM (w_tabname) INTO TABLE <t_itab>.
where w_tabname is the table name given by the user, and t_itab is field symbol.
but some requirement is like that i need to pick up only few fields, which user will give at the runtime.
I can take those fields from file into any internal table,
but the problem is that instead of select *, i need to put selected the fields given by user.
i have tried with field symbol, it is not working (as per my knowledge.)
do i need to create any structures dynamically ?
Can anybody please help in this regards ?
Your help will be surely rewarded with points.
Waiting for reply..
Regards
Vinay

Hi Vinay ,
Adding to the below code , you can use some more fields in the select stmt as below:-
REPORT ychatest.
PARAMETERS : p_field1 LIKE dd03l-fieldname,
 P_field2 LIKE dd03l-fieldname,
 p_table LIKE dd03l-tabname.
FIELD-SYMBOLS : <fs> TYPE STANDARD TABLE.
SELECT (p_field1) (p_field2) ( FROM (p_table) INTO TABLE <fs>.
Now p_field1 & p_field2 belong to the same table since you have give the user to enter only one table name.
This should work fine.
please try & let me know .
Thanks & Regards,
Daniel

How to select few columns based on certain conditions

i have a table with n columns, of which in 2 columns each may have similar values for certain number of rows
i want to find if one of the entries in a column has same value in the next few rows
For example the two columns are
supplier_name supplier_code
hll 013
hll 013
hll 013
hll 013
if i first encounter hll with 013 i have to check whether hll with 013 record is further available
if it is so i have to display the columns for which multiple entries is available
can anyone send the query

Duplicate thread?
selecting certain columns in oracle

Help in selecting required columns from duplicates records

HI,
I have a table like below
empid,ename,sal,Jdate,Rdate,deptid,designation,Terminated_Emp
1,a,100,15-May-2014,11,SE,NO
1,null,null,15-OCT-14,null,null,YES
I want output like below
1,a,100,15-May-2014,15-OCT-2014,11,SE,YES
Please help me in writing sql query, can we write case condition for this?
Naresh

HI
And I have to select remaing records from first row only
I mean
empid,ename,sal,Jdate,Rdate,deptid,designation,Terminated_Emp
1,a,100,15-May-2014,null,11,SE,NO
1,null,null,null,15-OCT-14,null,null,YES
2,b,200,12-Jan-2014,null,12,SSE,NO
I want records like
1,a,100,15-May-2014,15-OCT-2014,11,SE,YES
2,b,200,12-Jan-2014,null,12,SSE,NO
Naresh

Select a column from an inner select statement

On my oracle 10g running on solaris, I am running this little query that gives me
the right results (I needed one more column and not able to get to it, hence looking for some ideas here)
Select a.othtype, a.date_from, a.date_to,b.process
from bio_plan_other a,
bio b,
Where a.bkey = b.bkey
and a.date_from <
(Select min(f.edate)
from bio_service f
where f.bkey = a.bkey) ;
In the inner select table bio_service, I have an additional column called service_name. Hence for the min(f.edate), there is a corresponding service_name.
How do I get that service_name in my topmost SELECT statement ?
In other words, I wanted to do something like...
Select a.othtype, a.date_from, a.date_to,b.process,f.service_name
from bio_plan_other a,
bio b,
Where a.bkey = b.bkey
and a.date_from <
(Select min(f.edate)
from bio_service f
where f.bkey = a.bkey) ;
But this gives me,
ORA-00904 : f.service_name : invalid identifier
Thanks

Hi
Not knowing your datamodel, but you probably have more service names per date, so did you try this?
WITH t_bio_service AS (SELECT b2.bkey, b2.min_edate, b.service_name
 FROM bio_service b
 , (SELECT MIN(edate) min_edate, bkey
 FROM bio_service
 GROUP BY bkey) b2
 WHERE b2.bkey = b.bkey
 AND b2.min_edate = b.edate)
SELECT a.othtype, a.date_from, a.date_to,b.process, t.service_name
FROM bio_plan_other a,
 bio b,
 t_bio_service t
WHERE a.bkey = b.bkey
 AND t.BKEY = b.bkey
 AND a.date_from < t.min_edateHope this helps.
Regards Auke

Is There any difference in Set @Var=ColmnName and Select @Var =Column from Table. I am using it in SSIs for Dynamic Packages

hey there, Kindly elaborate me the difference Between Set and Select clause while Providing value to a Variable or Even a Column.
    Declare @var nvarchar(Max)
    Set @var= (Select TestName+ cAst(TestId as nvarchar) from TblTest where TestID=1)
    Insert into TblTest(TestName)
    Values (@var)
    Declare @var nvarchar(Max)
    Select @var= (Select TestNAme + Cast(TestId as Nvarchar(MAx)) from TblTest Where TestID=1)
    Insert into TblTest(TestName)
    Values(@var)
Values and result I am getting is Same in Both.
But just wondering if there is any Difference. and if Any which Should I prefer.
And Which one uses less Resources.

I like this article
http://vyaskn.tripod.com/differences_between_set_and_select.htm
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

Removing columns from the result of a query via SQL

Hi
I have written a SQL query that returns the following results:
TableA
A1.ColA A2.ColB
1234 567
567 1234
I would like the query not to display the second row since the first row already has the required info.
The SQL query is
Select A1.ColA, A1.ColB
from TableA A1,
TableA A2
where A1.ColA <> A2.ColB
As you can see I'm comparing the values in the same column.
Regards
Bilal

As I can see none of the previously mentioned solutions here work in the proper way, so you can try this:
SQL> with t as (select 1234 colA, 5678 colB from dual union all
2 select 5678 colA, 1234 colB from dual union all
3 select 2 colA, 1 colB from dual union all
4 select 1 colA, 3 colB from dual)
5 select colA, colB
6 from (select colA,
7 colB,
8 row_number() over(partition by least(colA, colB), greatest(colA, colB) order by colA) rn
9 from t)
10 where rn = 1
11 /
 COLA COLB
 2 1
 1 3
 1234 5678

Selecting only columns from a table

Could someone help me with a SQL query to select all the columns in a table as an array?

Hello
if you are on 9i upward, you can define a pl/sql table of the rowtype and use bulk collect to populate that table, otherwise, you can create separate array types for each column:
create table dt_test_tab (id number, col1 varchar2(10));
insert into dt_test_tab values(1, 'X');
insert into dt_test_tab values(2, 'X');
insert into dt_test_tab values(3, 'X');
insert into dt_test_tab values(4, 'X');
insert into dt_test_tab values(5, 'X');
commit;
declare
    type tt_TestTab is table of dt_test_tab%rowtype index by binary_integer;
    type tt_TestTabid IS TABLE OF dt_test_tab.id%type index by binary_integer;
    type tt_TestTabcol1 IS TABLE OF dt_test_tab.col1%type index by binary_integer;
    lt_TestTab     tt_TestTab;
    lt_Id          tt_TestTabId;
    lt_col1        tt_TestTabcol1;
begin
    --use the rowtype array
    select
         id,
         col1
    BULK COLLECT INTO
         lt_testtab
    FROM
        dt_test_tab;
   --use the individual column arrays
    select
         id,
         col1
    BULK COLLECT INTO
         lt_id,
         lt_col1
    FROM
        dt_test_tab;
    dbms_output.put_line('loop through the rowtype array');
    FOR x in lt_TestTab.FIRST..lt_TestTab.LAST LOOP
       dbms_output.put_line(to_char(lt_TestTab(x).id)||to_char(lt_TestTab(x).col1));
    end loop;
    dbms_output.put_line('loop through the two column arrays');
    FOR x in lt_Id.FIRST..lt_Id.LAST LOOP
       dbms_output.put_line(to_char(lt_Id(x))||to_char(lt_Col1(x)));
    end loop;
END;

Selecting fewer columns from cached results

Similar Messages

Maybe you are looking for