Help on group by clause

Question

I have a table which is defined like this.
unique_number NOT NULL VARCHAR2(64),
input_unique_number VARCHAR2(64),
source_system VARCHAR2(40),
end_system VARCHAR2(40),
type VARCHAR2(32),
date_time_sent TIMESTAMP(6),
transaction_name VARCHAR2(60)
type will have input, output and failed
For every input message a unique_number is generated it will have source_sytem value populated but not end_sytem value
For every output message a unique_number is generated plus it will have input_unique_number same as that of input message.
For every output message source_system and end_system values are populated.
I am trying to create a report which will give input, output and failed count for each end site, but output is getting the blank ids grouped for the input_messages
select
count(case type when 'input' then Unique_id end) as I,
count(case type when 'output' then Unique_id end) as O,
count(case type when 'failed' then Unique_id end ) as F,
transaction_name,
end_system
from message_table
where to_char(date_time_sent,'DD-MON-YY HH24:MI:SS') between '18-MAY-11 09%' and '19-MAY-11 11%'
and source_system='&Source_system'
group by transaction_name, end_system
ORDER BY end_system;
I got the following output
I     O          F     TRANSACTION_Name     END_SYSTEM
0     11          0     DistProdQualityEvent     EANCM5051256614152
22     0          11     DistProdQualityEvent     EANCM
0     1          0     DistInventoryReport     EANCM5051351029368
1     0          0     DistInventoryReport     EANCM
But I trying for output something like this
I     O          F     TRANSACTION_Name     END_SYSTEM
22     11          11     DistProdQualityEvent     EANCM5051256614152
1     1          0     DistInventoryReport     EANCM5051351029368
Can you please let me know is such an output is possible or not

Answer 1

Hi,
Perhaps this is what you want:
select    count(case type when 'input'  then Unique_id end) as I,
       count(case type when 'output' then Unique_id end) as O,
       count(case type when 'failed' then Unique_id end) as F,
       transaction_name,
       MAX (end_system)     AS max_end_system               -- Changed
from        message_table
where        date_time_sent     between TO_DATE ( '18-MAY-2011 09'     -- Compare DATEs, not strings
                                , 'DD-MON-YYYY HH24'
                    and      TO_DATE ( '19-MAY-2011 11%'
                                , 'DD-MON-YYYY HH24'
and       source_system          = '&Source_system'
group by  transaction_name     -- Without end_system               -- Changed
ORDER BY  max_end_system                                    -- Changed
;If you include end_system in the GROUP BY clause, then rows with end_system='EANCM5051256614152' will not be grouped with rows with end_system='EANCM'.
I hope this answers your question.
If not, post a little sample data (CREATE TABLE and INSERT statements, relevant columns only), and also post the results you want from that data.
Explain, using specific examples, how you get those results from that data. Why should the output say 'EANCM5051256614152' and not ''EANCM'?
Always say which version of Oracle you're using.
Edited by: Frank Kulash on Jun 3, 2011 8:49 AM