Aggregating a union query

OBIEE 10.1.3.3: We have a number of HR turnover reports that require calculations that use a formula of: #txns / average headcount. We need to run these reports for whatever beginning/ending dates the users enter. And, of course they need to be sliced and diced by any number of dimensions. I came up with a couple of different schemes to do this but keep running into roadblocks. It seems that a union is my best bet. here's the latest one generated:
SELECT saw_0, saw_1, saw_2, saw_3, saw_4, (saw_3 + saw_4)/2 saw_5,
(saw_2 / (saw_3 + saw_4)/2)*100 saw_6
FROM
((SELECT saw_0, saw_1, saw_2, saw_3, saw_4 FROM ... WHERE...)
UNION
(SELECT saw_0, saw_1, saw_2, saw_3, saw_4 FROM ... WHERE...)
UNION
(SELECT saw_0, saw_1, saw_2, saw_3, saw_4 FROM ... WHERE...)
)) t1 ORDER BY saw_0, saw_1
the first union select puts #txns in saw_2 and zeroes in saw_3 and saw_4
the second union select puts 1st date headcount in saw_3 and zeroes in saw_2 and saw_4
the third union select puts 2nd date headcount in saw_4, and zeroes in saw_2 and saw_3
The saw_6 calc won't work because the overall query needs to be aggregated. Here's the problem: When I add sum() clauses to the main query, it blows up because of no group by clause, and if I go into Advanced and try to add the group by and hit set sql, Answers cuts off the query and tells me there's no group by, even though it properly places the saw_0,saw_1 fields in the group by box below. Seems like a bug in Answers, because when I take the SQL I altered and run it from the Admin SQL tool, it returns the result I wanted.
Any ideas? If you got this far, thanks for reading all of this...
Greg

I am using very similar query with union all and sum () to get approval ratings between any given dates, but I did not explicitly specify groupby and it works. I am using 7.8.4.

Similar Messages

How to get Average of a column in UNION query

Hi All,
I will try to explain the issue as much as I can and If you need clarification on any piece please let me know.
So, I have a Union Query with two sets of Criteria.
The first Criteria Brings the list of Buildings and a set of Measures for it. The second criteria does the same but with a different set of filters.
No I use a UNION between these both criteria and present them in a Pivot view so that we only have one row per each building. The values in each of the criteria in the UNION query are Summed
Ex:
Building
Metric 1
Metric 2
1
20
30
1
25
35
2
40
50
2
45
55
So as I show in the pivot the result is
Building
Metric 1
Metric 2
1
45
65
2
85
105
Now the issue is with the Grand Total.
I want the Grand Total to be an Average of the Buildings. So for Metric 1 The Grand Total should be (45+85)/2 because we have two buildings so the answer should be 65.
How can I get that.
Bottom Line: Need an Average as the Grand Total when the Aggregation rule on the column is set to Sum.
we use 11.1.6.10 version

Hi VJ
What you are asking is impossible in the pivot table itself as you are asking one part of the pivot table to aggregate via one method, and another to aggregate via another method, on the same column ..
It can be done using a calculated item. For your example above you'd use the formula ( $1 + $2) / 2.
.. But thats HIGHLY explicit, and would only work if you KNOW how many buildings will come back in the query.. EVERYTIME.
Why not restructure the query to combine it into a single query using an combinations of filters with OR in between?

Sum of columns in union query

hi, i am using oracle 10g database..
how to get the sum of column in union query ?
     select * from (select 100 records from dual), (select 50 available from dual)
union
select * from (select 200 records from dual), (select 150 available from dual)
display should be like
                 records         available
                  100               50
                  200               150
total            300               200thanks ...

Peter vd Zwan wrote:
try this:Grouping by records will not produce correct results:
SQL> with a as
2 (
3 select * from (select 100 records from dual), (select 50 available from dual)
4 union
5 select * from (select 100 records from dual), (select 100 available from dual)
6 union
7 select * from (select 200 records from dual), (select 150 available from dual)
8 )
9 select
10    case when grouping(records) = 0 then null else 'total' end tot
11    ,sum(records)   records
12    ,sum(available) available
13 from
14    a
15 group by
16 rollup (records)
17 ;
TOT      RECORDS AVAILABLE
             200        150
             200        150
total        400        300
SQL> select case grouping(rownum)
2     when 1 then 'Total'
3   end display,
4          sum(records) records,
5          sum(available) available
6    from (
7            select * from (select 100 records from dual), (select 50 available from dual)
8           union
9            select * from (select 100 records from dual), (select 100 available from dual)
10           union
11            select * from (select 200 records from dual), (select 150 available from dual)
12          )
13    group by rollup(rownum)
14 /
DISPL    RECORDS AVAILABLE
             100         50
             100        100
             200        150
Total        400        300
SQL> SY.

Parameter in union query

Hi Guys,
I want to know how to create parameter in case of union query in 8.8. with new token option.
Thanks Rashid

Hi Rashid,
It seems you are asking question regarding Crystal Report. Is that true?
Thanks,
Gordon

View using Union Query

All,
How to create a view using a union query in sql developer. I tried doing but when i check the syntax I got an error
"The SQL syntax is valid, however the query is invalid or uses functionality that is not supported.
Declarative query support does not currently include UNION, INTERSECT or MINUS"
Could some one help me out
Regards

792011 wrote:
All,
How to create a view using a union query in sql developer. I tried doing but when i check the syntax I got an error
"The SQL syntax is valid, however the query is invalid or uses functionality that is not supported.
Declarative query support does not currently include UNION, INTERSECT or MINUS"
Could some one help me outuse different/upgraded client since UNION is valid SQL

Prompts for Combine similar request (Union query)

I have been given 5 tables for creating one report out of each table. I created aliases and
necessary self joins. I now need a summaryreport that combines these five tables. I tried the
"combine similar request" and the report functions fine. Now I have to create two prompts, one for
country and one for job. These are fields available in 4 tables and have the same format alongwith
the same datatypes. Please help me with this.
Table 1 Report A
Table 2 Report B
Table 3 Report C
Table 4 Report D
Table 5 Report E
Table 1 - Table 4 contains Country and Job field which is of the same format and datatype.
Table 5 does not have these and its not necessary for us to prompt this.
This has no star schema setup and I am not in a position to get it done as well. I considered each
table to be a fact table and created aliases, selfjoins etc. Please help with the procedure to
create the two dashboard prompts. If I choose USA then the first four amongst the union queries
should reflect data for the USA. The same applies for Job.

Hello
Thanks for the reply.
My problem here is that I have 5 different tables with country fields in them. They are of the same format and datatype but there is no join between these tables. If I choose one Country field from any particular table, then the combine similar request doesnt seem to take it for other queries where the fields come from a different table.
eg.
Table 1 - table 5 have country fields
I choose to create a prompt based on Country from Table 1. The union query doesnt seem to reflect on filtering values on the queries from other tables. It would filter only the values from table 1.
Please let me know as to what can be done!

Alternative/enhancement to union query? Duplicate - please delete

Hello,
I have 2 sets of similar data which I am joining together with a union query.
Simplified E.g.
*Table A*
Name Dept
Ben    1
Amy   3
Ken    3
*Table B*
Name Dept
Jim    1
John 2
Sue   3
select Name, Dept
from table A
UNION
select Name, Dept
from table BWhat I would like to do is ONLY include data from the second part of the union query where it joins with some data from the first part of the union query.
E.g.
select Name, Dept
from table A
UNION
select Name, Dept
from table B
where Dept in (select distinct Dept from A)Therefore John that works in Dept 2 would not be included in the results.
Is there a better or more efficient way of doing this than my example using the subquery as part of the where clause?
Many thanks,
Hazel
Edited by: user11979866 on 04-Nov-2010 08:48

Hazel:
It would really depend on the sizes of the tables and the indexes available, but that is probably the way I would write it, at least as a first cut.
You could lose the distinct in the in sub-query since it is implied by the in itself, and Oracle will only return one row from the sub-query for each value of dept in A.
John

Alternative/enhancement to union query?

Hello,
I have 2 sets of similar data which I am joining together with a union query.
Simplified E.g.
*Table A*
Name Dept
Ben    1
Amy   3
Ken    3
*Table B*
Name Dept
Jim    1
John 2
Sue   3
select Name, Dept
from table A
UNION
select Name, Dept
from table BWhat I would like to do is ONLY include data from the second part of the union query where it joins with some data from the first part of the union query.
E.g.
select Name, Dept
from table A
UNION
select Name, Dept
from table B
where Dept in (select distinct Dept from A)Therefore John that works in Dept 2 would not be included in the results.
Is there a better or more efficient way of doing this than my example using the subquery as part of the where clause?
Many thanks,
Hazel

Hi, Hazel,
What you posted is as good as you can get without changing the tables.
You don't have to say DISTINCT in the IN-subquery:
where   Dept in ( select Dept
                   from    A
                )will get the same results, perhaps a little faster.

Customised sub-total via UNION query - how to exclude from totals?

Hi,
I have a client requirement that a couple of companies have their own sub-total in a report, a treatment which applies to no other companies in their group.
I have achieved this via a union all, but they also want a total of companies.
Is there anyway to emulate a Grand Total, other than another union query such that the double count is not double counted?
thanks for any ideas.
Robert.

only solution which i can think of is to create another union request for the grand total.. this way you wont add up your sub totals twice..

Dynamic Aggregation Using BI Query Wizard

Hi,
Is it poosible to perform dynamic aggregation using Oralce Query Wizard. In other words , is it possible to change/filter the dimension members of a particular dimension and recalculate the aggregation with out changing the fact/dimension table contents.
Your help is appreciated.
Panneer

Hi,
Not sure exactly what you are tying to achieve. However, if your data is stored in Analytic Workspace objects then you can recalculate data via OLAP DML programs or models dynamically from within a BI Beans application. The AW environment is always attached in read-only mode as a default. Hence, you can change the data during a session and then dump the changes when the user exits the application.
From within a relational environment it is very difficult to create the same approach since it is possible a commit could be issued. Unless you mark the required tablespaces as read-only. In which case you could execute PL/SQL procedures to recalculate data. However, not sure how you would resolve the refresh of any related MVs
Hope this helps
Keith Laker
Product Manager
Oracle Business Intelligence Beans

How do I make a 'union' query?

Hi,
How do I make a union query in JDev? Something like this.
select 'x' from dual
union
select 'y' from dual;
to get a rowsetinfo for showing in a combobox control? Don't ask
me to create view in the database:-). It may not be practical for
all such adhoc queries.
--Gopal
null

Hi
Sorry for that one. I didn't do my rtfm before asking that. There
is a constructor for QueryInfo which takes any sql statement.
--Gopal
Gopal (guest) wrote:
: Hi,
: How do I make a union query in JDev? Something like this.
: select 'x' from dual
: union
: select 'y' from dual;
: to get a rowsetinfo for showing in a combobox control? Don't
ask
: me to create view in the database:-). It may not be practical
for
: all such adhoc queries.
: --Gopal
null

Create a Union query that automatically adds/contains new tables with a similar name

Dear All,
For a research project i am upscaling from an Access database to SQL server/ SQL azure since the database got far too large. Now i am quite well known with Access, but to be quite honest, I am a SQL rookey. So a lot to learn for me here.
what I need is a union query that automatically adds new research dumps to the query. All research dumps have a table name starting with TD_EXP (Table Dump _ Export). In Access VBA I simply made a loop that builds up the SQL statement, but i have no clue
how to do this within SQL Server...
So, I need the UNION Query to select * from TD_Exp_2012_06_06_15_48 UNON all the TD_Exp_XXXX_XX_XX_XX_XX that are added in the database,
Who could help me out please with a dynamic SQL statement.
Best regards Jurjen

you can use INFORMATION_SCHEMA.TABLES view for that
do like below
DECLARE @Query varchar(max)
SELECT @Query = STUFF((SELECT ' UNION ALL ' + 'SELECT * FROM ' + TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME LIKE 'TD_EXP%' FOR XML PATH('')),1,11,'')
SELECT @Query
Add one more condition as below if you want to exclude views from above
DECLARE @Query varchar(max)
SELECT @Query = STUFF((SELECT ' UNION ALL ' + 'SELECT * FROM ' + TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME LIKE 'TD_EXP%' AND TABLE_TYPE = 'BASE TABLE' FOR XML PATH('')),1,11,'')
SELECT @Query
Please Mark This As Answer if it helps to solve the issue Visakh ---------------------------- http://visakhm.blogspot.com/ https://www.facebook.com/VmBlogs

Trying to narrow down a UNIONed query

I've compared 2 tables using MINUS. There should have been no difference. Since there was, I used UNION to display the corresponding rows between both tables. Now, in my result set, I have to search through many rows that don't have correspondents to find rows that do.
Basically how do I narrow the UNION query down so that only the corresponding pair rows that have differences in their columns show up? I'm trying to use the primary key which is a combination of the first 5 columns. My query looks like this:
select
TRADE_DATE, ORDER_ROUTE_UID, ORDER_ENTRY_DATE, LEG_NBR, GROUP_SEQ_NBR,
AFFL_FIRM_GROUP_ID, REC_CREAT_DATE
from ors_firm_affl_hist@MSS_JOBS_HISTP01_LINK01.WORLD
union
select
TRADE_DATE, ORDER_ROUTE_UID, ORDER_ENTRY_DATE, LEG_NBR, GROUP_SEQ_NBR,
AFFL_FIRM_GROUP_ID, REC_CREAT_DATE
from ors_firm_affl_hist;
How can I use TRADE_DATE, ORDER_ROUTE_UID, ORDER_ENTRY_DATE, LEG_NBR, GROUP_SEQ_NBR in order to have the query display only the rows that have correspondents in the other table as opposed to rows that just aren't in either table?

Not sure if this meets your requirements;
SQL> with t1 as (
   select 'jon' first_name, 'smith' last_name, 'j' middle_init from dual union all
   select 'tom', 'a', 'a' from dual union all
   select 'bob', 's', 's' from dual union all
   select 'Stan', 'm', 'l' from dual union all
   select 'Kyle', 'b', 'r' from dual union all
   select 'Kenny', 'm', 'r' from dual union all
   select 'Eric', 'c', 'c' from dual),
     t2 as (
   select 'jon' first_name, 'smith' last_name, 'p' middle_init from dual union all
   select 'tom', 'a', 'l' from dual union all
   select 'bob', 's', 'j' from dual union all
   select 'Stan', 'm', 'l' from dual union all
   select 'Kyle', 'b', 'r' from dual union all
   select 'Kenny', 'm', 'r' from dual union all
   select 'Eric', 'c', 'c' from dual)
   select 'T1' source, first_name, last_name, middle_init from t1
   minus
   select 'T1' source, first_name, last_name, middle_init from t2
union all
   select 'T2' source, first_name, last_name, middle_init from t2
   minus
   select 'T2' source, first_name, last_name, middle_init from t1
order by 2,3,1
SO FIRST LAST_ M
T1 bob   s     s
T2 bob   s     j
T1 jon   smith j
T2 jon   smith p
T1 tom   a     a
T2 tom   a     l
6 rows selected.

Union query optimalization

Help me please. I have this query with a Union
select p.cdpwet, p.jaar, p.vnrproc
from abgrp p,tbsltbetro bbt
where p.vnrbslt = bbt.vnrbslt
and bbt.cdsrt like :n_rol
and bbt.nrbetr in (select i.nrbetr from tuser i
where i.naam like :n_naam
and (i.cdsrtrpers like :n_cdsrtrpers
or :n_cdsrtrpers is null)
and :n_nrbetr is null
and i.indsrtpers = 'R'
union
select to_number(:n_nrbetr)
from dual
where :n_nrbetr is not null)
I will the same result but without a Union.
Regards,
Patrick

Hi Patric,
probably if you dont want to display UNIQUE values and you are sure that both parts of your query will give different o/p and you just want to club them.. Use UNION ALL.. Your performance problem will be solved, especially when the data fetched is large. Union finds the two data sets and try to sort them to display UNIQUE records only where as UNION ALL just merges them so saving time on sorting.
Hope that this will help you.
Regards
Santa

Can we change filter column values in a union query using Dashboard prompt

Hi,
I have a requirement in which have to draw a chart report for last six months?.
When i query against database, my query got different where clause for open month and close month.
Say for july month bar, it got close month as july and open month as july, june, may. Like this, for june month bar, it has to have close month as june and open month as june, may, april. like this, i created six union queries using repository variable.
Want to give control using dashboard prompt for close month and open month?
my query look like this.
select close month1, fact1 from table1
where close month = july and open month in (july, june, may)
UNION
select close month2, fact2 from table1
where close month = june and open month in (june, may, april)
UNION
select close month3, fact3 from table1
where close month = may and open month in (may, april, march)
UNION
select close month4, fact4 from table1
where close month = april and open month in (april, march, february)
UNION
select close month5, fact5 from table1
where close month = march and open month in (march, february, january)
UNION
select close month6, fact6 from table1
where close month = february and open month in (february, january, december)
Welcome your suggestions on this?.
Thanks

Hi Karol,
Yes the prompt is working fine when i put this following method under the "Button" Component.
APPLICATION.openPromptDialog(800, 600);
Somehow the when i use the "Input Field" and use the "Button" Component and write this following logic inside it i don't see the prompt is not taking the values what i am entering in input field i assume that its issue with data-source not getting refreshed.
Variable=INPUTFIELD_1.getValue();
APPLICATION.setVariableValue("PercentageIncrease",Variable);
Thanks,
Kumar

Aggregating a union query

Similar Messages

Maybe you are looking for