Help finding the Max Total

Hi,

I have the following code

SELECT
PR.WBS2,
SUM(CASE WHEN LedgerAR.Period = '200408' AND LedgerAR.TransType <> 'CR'
THEN Ledgerar.amount * - 1
ELSE '0' END) AS BillExt
FROM PR
LEFT JOIN Ledgerar ON PR.WBS1 = Ledgerar.WBS1 AND
PR.WBS2 = Ledgerar.WBS2 AND PR.WBS3 = Ledgerar.WBS3
WHERE PR.WBS2 <> '98' AND pr.wbs2 <> '9001'
AND pr.wbs2 <> 'zzz' AND pr.wbs3 <> 'zzz' AND
pr.wbs1 = '001-298'
GROUP BY PR.WBS2

It prints out:
Wbs2 BillExt
0141 0
0143 0
1217 20580

I want the code to return the wbs2 code '1217' because it has the highest amount in BillExt '20580'.

Can someone help me with this?

Thanks.
lauraThe quick and dirty version...

select top 1 a.WBS2, a.BillExt
from
(SELECT
PR.WBS2,
SUM(CASE WHEN LedgerAR.Period = '200408' AND LedgerAR.TransType <> 'CR'
THEN Ledgerar.amount * - 1
ELSE '0' END) AS BillExt
FROM PR
LEFT JOIN Ledgerar ON PR.WBS1 = Ledgerar.WBS1 AND
PR.WBS2 = Ledgerar.WBS2 AND PR.WBS3 = Ledgerar.WBS3
WHERE PR.WBS2 <> '98' AND pr.wbs2 <> '9001'
AND pr.wbs2 <> 'zzz' AND pr.wbs3 <> 'zzz' AND
pr.wbs1 = '001-298'
GROUP BY PR.WBS2) a
order by BillExt desc|||Nothing dirty about it...|||That would be nice if there was only 1 record that needed to be returned.
select a.WBS2, BillExt=max(a.BillExt)
from
(SELECT
PR.WBS2,
SUM(CASE WHEN LedgerAR.Period = '200408' AND LedgerAR.TransType <> 'CR'
THEN Ledgerar.amount * - 1
ELSE '0' END) AS BillExt
FROM PR
LEFT JOIN Ledgerar ON PR.WBS1 = Ledgerar.WBS1 AND
PR.WBS2 = Ledgerar.WBS2 AND PR.WBS3 = Ledgerar.WBS3
WHERE PR.WBS2 <> '98' AND pr.wbs2 <> '9001'
AND pr.wbs2 <> 'zzz' AND pr.wbs3 <> 'zzz' AND
pr.wbs1 = '001-298'
GROUP BY PR.WBS2) a
group by a.WBS2
order by BillExt desc|||Now going a little further once I find the maximum total what if I then have to insert a value in that record. Is there a way to do this without using subqueries and instead using case statements?

For instance :

Wbs2 BillExt MaxValue
0141 0 0
0143 0 0
1217 20580 1

Thanks,
Laura|||That would be nice if there was only 1 record that needed to be returned.

Which is wat she asked for...I left my mind reading hat at home...

Laura....INSERT What to Where?

Insert into that record?

I don't understand.|||I'm sorry it's hard to describe what I am asking.

If I have the following information in a table:

WBS1 Amount MAX
0141 0
1217 2
1222 200

I first want to find the maximum of amount which is 200. Next, I want to put a 1 in the MAX column where the largest sum appears in amount.

Ending result:

WBS1 Amount MAX
0141 0
1217 2
1222 200 1

Is this possible?|||Well, once you know the ID of the record with the highest value you can issue and UPDATE statement to set its MAX field value to 1. But you will also need to issue an UPDATE statement to reset all other MAX field values that might have been the largest value before.

I can't recommend marking a record as the "Maximum value" when that status can change at any time. It is better to have a function or view or stored procedure that finds the maximum record dynamically.

Why do you want to mark this record and what are you going to do with it?|||Well, I simplified this example quite a bit to generate ideas for myself. But What I am supposed to do is create a report for accounting.

Accounting bills there clients based on services provided. Each service is designated codes. What is supposed to happen in one of the reports is to print the total reimbursable amount into the service that was billed the most.

There is no insertion are update that can be done it is just for display purposes only.

One of the restrictions that I have been under is that I cannot use subqueries only case statements which makes it even more difficult.

So I may have to go an entirely different path.

Thanks for your help,
Laura|||You can't use subqueries? What kind of lunacy is that?

You could still use a stored proc that first loads the MAX value's primary key into a variable and then uses the variable in subsequent queries.

Can't use subqueries? Kindly direct the person who gave you that directive to this informative website:

http://www.hov-hov.dk/you.htm|||Well, I simplified this example quite a bit to generate ideas for myself. But What I am supposed to do is create a report for accounting.

Accounting bills there clients based on services provided. Each service is designated codes. What is supposed to happen in one of the reports is to print the total reimbursable amount into the service that was billed the most.

There is no insertion are update that can be done it is just for display purposes only.

One of the restrictions that I have been under is that I cannot use subqueries only case statements which makes it even more difficult.

So I may have to go an entirely different path.

Thanks for your help,
LauraCan you "cheat" and JOIN a virtual table? Technically that isn't a sub-query.

-PatP

Side note to Blindman, I'm going to have to bookmark that site!

Help building a #temp table

I have the following query:

--build query for #temp table
select 'L' as RECTYPE, identity(int, 1, 1) as JELINENO, max(i.jeno) + 1 as JENO, 'AJE' as JECODE,
--build the fiscal year, when month = October, November or December add 1 to actual year
FY = case
when datepart(month, GetDate()) = '10' then datepart(year, GetDate()) + 1
when datepart(month, GetDate()) = '11' then datepart(year, GetDate()) + 1
when datepart(month, GetDate()) = '12' then datepart(year, GetDate()) + 1
else datepart(year, GetDate())
end,
-- The PD starts in April and increases by 1 each month
PD = case
when datepart(month, GetDate()) = '1' then '10'
when datepart(month, GetDate()) = '2' then '11'
when datepart(month, GetDate()) = '3' then '12'
when datepart(month, GetDate()) = '4' then '1'
when datepart(month, GetDate()) = '5' then '2'
when datepart(month, GetDate()) = '6' then '3'
when datepart(month, GetDate()) = '7' then '4'
when datepart(month, GetDate()) = '8' then '5'
when datepart(month, GetDate()) = '9' then '6'
when datepart(month, GetDate()) = '10' then '7'
when datepart(month, GetDate()) = '11' then '8'
when datepart(month, GetDate()) = '12' then '9'
else 'error'
end,
'1' as SUBPD, 'N' as REVFLAG, 'Journal Entry from MAXIMO' as HEADDESC,
--If the issuetype is a return the AMOUNT is a negative value, multiply by -1
AMOUNT = case
when m.linecost < 0 then m.linecost * -1
else m.linecost
end
into #temp
from matusetrans m, issuesreturns i
go
select * from #temp
drop table #temp

When I run this query, I get the following error:

Server: Msg 8118, Level 16, State 1, Line 2
Column 'm.linecost' is invalid in the select list because it is not contained in an aggregate function and there is no GROUP BY clause.
Server: Msg 8118, Level 16, State 1, Line 2
Column 'm.linecost' is invalid in the select list because it is not contained in an aggregate function and there is no GROUP BY clause.
Server: Msg 8118, Level 16, State 1, Line 2
Column 'm.linecost' is invalid in the select list because it is not contained in an aggregate function and there is no GROUP BY clause.
Server: Msg 208, Level 16, State 1, Line 1
Invalid object name '#temp'.

I can find no reason for this error. If I run the query without the AMOUNT = case statement, it works fine. Can anyone help?

Thanks,

Dave

The mistake only arises in the AMOUNT=CASE statement since it's the only one (except from the MAX(i.jeno) column) that's accesing a real column in the database. The rest of the columns are calculated.

So, as you already have an aggregate function (the MAX) the rest of the columns from real tables must either have an aggregate function or the SELECT clause must contain a GROUP BY clause.

You can't form a SELECT with only one MAX column and the rest of them normal columns without telling SQL how it should group the results from where to choose the MAX value from.

If you really want to achieve that you have to calculate the MAX column either before in a variable or in the same SELECT with a sub-SELECT query.

IE: SELECT var1, var2, var3, (SELECT MAX(othervar) FROM table)

Hope i was clear.
Good luck,
Alan.

|||

First of all I would compress the code and rewrite it as following.

Please post DDLs as it's not clear what do you want to do. the problem is clear as you have a reference to the column m.linecost that is not in a group by list. Just to get rid of the error you can add a group by abs(m.linecost), but it doesn't look like what you want to do.

select
'L' as RECTYPE,
identity(int, 1, 1) as JELINENO,
max(i.jeno) + 1 as JENO,
'AJE' as JECODE,

--build the fiscal year, when month = October, November or December add 1 to actual year

FY = datepart(year, GetDate()) + (datepart(month, GetDate()) + 2) / 12,

-- The PD starts in April and increases by 1 each month

PD = (datepart(month, GetDate()) + 8) % 12 + 1,

'1' as SUBPD,

'N' as REVFLAG,

'Journal Entry from MAXIMO' as HEADDESC,

--If the issuetype is a return the AMOUNT is a negative value, multiply by -1

AMOUNT = abs(m.linecost)

into #temp

from matusetrans m, issuesreturns i

|||Dave -

You can't use an Aggregate Function on only one column without using a Group By for the other columns

So you either need to group by m.linecost OR put an Agrregate around it - like max() or min().

To simpify your query

select
max(i.jeno) + 1 as JENO, -- YOU USED MAX HERE
AMOUNT = case
when m.linecost < 0 then m.linecost * -1
-- SO YOU EITHER NEED TO USE AN AGRREGRATE HERE OR
else m.linecost
end
into #temp
from matusetrans m, issuesreturns i
-- OR GROUP BY m.linecost HERE

Hope that helps

AWAL

Help - Finding the newest date without using max()

Hi,

Can anyone help please?

select notefield, modifiedon
FROM Table1
WHERE id = '100426' and
(statusfield like '%criteria1%' OR
statusfield like '%criteria2%')

Produces a list of records based upon the criteria. I would like to be
able to only show the newest dated record, from the modifiedon field.

I've tried max(modified) on, but as I am using an aggregate function in
the query I have to use GROUP BY, which notefield does not like as this
is a ntext field.

(I get:

Server: Msg 306, Level 16, State 2, Line 1
The text, ntext, and image data types cannot be compared or sorted,
except when using IS NULL or LIKE operator.
)

Any ideas please?select TOP 1 notefield, modifiedon
FROM Table1
WHERE id = '100426' and
(statusfield like '%criteria1%' OR
statusfield like '%criteria2%')
ORDER BY modifiedon DESC

--------
Alexander Kuznetsov
http://sqlserver-tips.blogspot.com/