Question

I have a tblMachineReports with the columns: Status(varchar),LogDate(datetime),Category(varchar), and MachineID(int).

I want to retrieve the latest status update from each category for every machine, so in effect getting a snapshot of the latest statuses of all the machines unique to their MachineID.

The table data would look like

Category – Status – MachineID – LogDate
cata – status1 – 001 – date1
cata – status2 – 002 – date2
catb – status3 – 001 – date2
catc – status2 – 002 – date4
cata – status3 – 001 – date5
catc – status1 – 001 – date6
catb – status2 – 001 – date7
cata – status2 – 002 – date8
catb – status2 – 002 – date9
catc – status2 – 001 – date10

Restated, I have multiple machines reporting on multiple statuses in this tblMachineReports. All the rows are created through inserts, so their will obviously be duplicate entries for machines as new statuses come in. None of the columns can be predicted, so I can’t do any =’some hard coded string’ comparisons in any part of the select statement.

For the sample table I provided, the desired results would look like:

Category – Status – MachineID – LogDate
catc – status2 – 002 – date4
cata – status3 – 001 – date5
catb – status2 – 001 – date7
cata – status2 – 002 – date8
catb – status2 – 002 – date9
catc – status2 – 001 – date10

What would the select statement look like to achieve this, getting the latest status for each category on each machine, using MS SQL Server 2008? I have tried different combinations of subqueries combined with aggregate MAX(LogDates)’s, along with joins, group bys, distincts, and what-not, but have yet to find a working solution.

Answer 1

Appologies if you have already tried this but this form of grouping and aggregation should work

select *
from tblMachineReports t
where LogDate = (select max(LogDate) from tblMachineReports t2 where t.machineID = t2.machineID and t.category = t2.category)

Try that, might need some fine tuning