List item
Hi I am trying to count the number of times an event is raised within the last 14 days of a date. simply put if i have an opened case on the 23rd of last month i want to check the previous 14 days of the 23rd to see if there have been another event raised of the same type within that period.
Is there a reason you can't do this in the SQL, and need to have it within the RDL?
The SQL should be easy, try making use of subqueries to get around the aggregate-of-an-aggregate hurdle.
If you need to do it within the RDL, the Lookup() and LookupSet() functions might make your life easier.
If this doesn't help you, could you post a bit more detail? Data structure may help.
Related
Recently I created an automated production scheduling tool through Excel that assigns a rank to items being produced in the same process, and then uses that rank in combination with the workload to create a schedule.
It functions exactly the way it is intended to, but due to the large amount of data and it being excel it has very slow performance, which is why I am looking to move the calculations over to SQL.
The general logic is like this:
-Always produce everything from the first day before the second day
-Always produce items from an earlier rank before items from a later rank
You can see how this plays out in the image below, where the line has 21.5 hours today, so items will be produced on day 1 until it equals 21.5, where the remainder is then carried over to day 2 and so on.
I was able to do this in excel using lengthy positional based formulas, but I am trying to think of a way to get the same result in SQL without having to rely on looking at the row above.
I am not sure how to convey something like 'Subtract from the available time production time of higher priority items produced on the same day'.
I apologize if the question is unclear, but any advice would be appreciated.
Image of Production Hours Cascading by Priority and Day
Example of Position-Based Fomula
Thanks to shawnt00, that put me in the right direction. Ultimately I had to modify the case statements a bit to go off of the cumulative total instead, but I was able to get the desired results using a sum() Over (partition by order by ) statement.
I have a view that lists employee (EmpID), request number (ReqNo), date request was opened (OpenDate) and the date it was moved to the next step in the process (AssignDate). What I am trying to do is get an average of the daily queue size. If EmpID 001 has 20 requests on 1/1/13, then has 24 on 1/2/13, 21 on 1/3/13 the average over 3 days should be 21.66, rounded up to 22. I have the following view:
CREATE VIEW EmpReqs
AS
SELECT [EmpID], [OpenDate], [AssignDate], [ReqID]
FROM [Metrics].[dbo].[Assignments]
WHERE OpenDate BETWEEN '01/01/2013' AND '12/31/2013' AND
[EmpID] IS NOT NULL AND
[ReqNo] NOT LIKE 'M%'
I then wrote a query to pull individual employee's queues per day:
/* First attempt to generate daily queue #s */
SELECT * FROM BLReqs
WHERE [BusLiaison] LIKE 'PN' AND
[OpenDate] <= '11/15/2013' AND
[AssignDate] > '11/15/2013'
Because no one has attempted to pull this information before, I have no way of verifying how accurate the above is. I tried using current dates, since I can see those in our database to compare but the code doesn't work, nothing is returned when I change the dates to 2014 and run my query.
What is the easiest way to verify that my code is correct, short of manually counting a day's queue?
Can anyone see any issues with the above scripts?
Is there a way to get the above code to work with current dates?
This question is really hard to answer because it is kind of broad and has little information at the same time. I'll try anyway:
Because no one has attempted to pull this information before, I have
no way of verifying how accurate the above is.
Try checking the result of this query for a few sampled dates.
I tried using current dates, since I can see those in our database to
compare but the code doesn't work, nothing is returned when I change
the dates to 2014 and run my query.
So clearly, the query is not working. You should probably find out why. Run the query for a date of which you know that it should return results but doesn't. Remove conditions one by one to see which one is incorrectly removing all rows. This should be enough to identify the bug.
Can anyone see any issues with the above scripts?
No, looks fine. A very simple query. That's why I said that we have too little information. There is some key piece of information missing that allows us to find the bug.
Is there a way to get the above code to work with current dates?
Stop staring at the code and hoping for a revelation. Debug it. Experiment.
I am very new to BO webintelligence.
I am doing a very simple query, retrieve Sales Amount for dates between 2012 and 2013. Just for this simple query when I run it, my BO crashes or gets stuck on "please wait" window. Why is this happening? If I select like 3 or 4 days like between Jan 1st 2012 and Jan 4th 2012, it runs fine. Is there anything from my end that I am doing wrong? This is in production mode.
I also wanted to point out that I have tried limiting my data set for specific region etc.
Too many unknowns to successfully assist. I presume this is webi 3.1.x or 4.1.x and I presume that the Sales Amount is a measure and that your query includes a date which does not include a time component. If you only have these two objects in your results pane and have the region in your conditions then pulling something like [Date] and sum([Sales Amt]) should not take long to execute. As previous poster suggests, try to execute the sql in a tool like SQL*Developer or Management Studio or however the source database / OLAP is stored.
Even though you've limited the data to a "region" this may be too much data. Try selecting a smaller result set of 100 rows or changing the option for "retrieve duplicate rows" on the query panel.
If possible, post the query from your report using generic object names?
I need to create a SSRS report where the user specifies the start and end date (e.g. March 2012 to July 2012) and the result is a report with the following headings:
User Department Product Mar-12 Apr-12 May-12 Jun-12 Jul-12
The dates are stored in a datetime column (e.g.2012-06-11 14:48:04.787) in my SQL table which I'm using to build the time-series report.
Do I need to convert/transpose my SQL table first? Or should I be processing this in SSRS?
Your advice is much appreciated!
Based on your question it appears all your dates are in a single datetime column and your trying to display them, formatted as a month, in a single report with the report columns grouped and renamed based on the month. This is a pivot and I don't think you can do that within SSRS. The dataset will already need to be pivoted by the time SSRS gets it.
So, to me this sounds like something you'll have to do both pre and post SSRS. I did a similar report using days of the month. You can see the winning answer here. as opposed to months of the year.
Thinking through your logice, I think you'll probably have to build the query with the pivot and then send to SSRS. On the SSRS side you need 12 columns for all 12 months and then put some logic to hide the ones that don't get data. Your original pivot will need to be smart enough to pass the empty months so SSRS doesn't error out. The end result will appear the same but needs a lot of pre-thought and design effort up front.
You do the processing using SSRS - you would design it in the GUI editor(BIDS ) , and you can customize things using the VBasic code also(i.e, dates and logic)
Looks like you need an intro, I really recommend this video
Wonder if anyone has done this before and what solution you came up with?
I have a Stored Proc that basically is a financial aged debtors query. I have added params to it so that you can select the Start Date, Frequency and Period length of the query. e.g. I could pick 12 months back from TODAY or 01/01/2012 or 4 Quarters back or 1 year back depending on the detail of the output I wish to see.
All well and good and when I run it in SQL it is lovely and works beautifully. I have pivoted the data within the SP in different ways depending on the Period param so that the columns show correctly (i.e. Shows the month name column for months, Quarter name for quarters etc etc.) so the pivot is conditional on the input parameter.
As I say in SQL it is a beautiful thing... now how the hell do I display it in SSRS!? :o)
As the PIVOT is conditional I can't see the available fields and I've got to a point where I can't think anymore of a way around it.
Any help appreciated!
Assuming I got you right, there is an option to place the entire SP in the Dataset.
In the dataset properties , change the query type to Store Procedure.
Hope it helps for you...