Statistical Median with T-SQL (2012)

I have been interacting with my statistician colleagues lately and there was a reporting requirement to calculate medians over a spread of continuously distributed values. Thankfully with SQL Server 2012, you do not have to use the ROW_NUMBER() function as you had to with earlier T-SQL versions.

PERCENTILE_CONT and PERCENTILE_DISC are now available in T-SQL 2012. According to the SQL BOL, these functions are described as:

PERCENTILE_CONT: Calculates a percentile based on a continuous distribution of the column value in SQL Server 2012. The result is interpolated and might not be equal to any of the specific values in the column.

PERCENTILE_DISC: Computes a specific percentile for sorted values in an entire rowset or within distinct partitions of a rowset in SQL Server 2012. PERCENTILE_DISC calculates the percentile based on a discrete distribution of the column values; the result is equal to a specific value in the column.

Let us put these functions to test and understand them a bit more in detail:

I have a sample table(ServiceCost) with ServiceIDs (that will be the data element to group by) and Cost (the field for which medians have to be calculated).


I calculated the medians using PERCENTILE_CONT and PERCENTILE_DISC as below:




              OVER (PARTITION BY ServiceID) AS MedianValue,


              OVER (PARTITION BY ServiceID) AS MedianDisc

  FROM ServiceCost

  order by ServiceID<code>


The result of this query is:


The difference between the way these two functions work is clear from the output for ServiceID: SA456. The PERCENTILE_CONT function calculated the mean of the values {62,73} to result in 67.5; while the PERCENTILE_DISC function resulted in the value 62 which is one of the column values in the data set.

We can now show off the power of T-SQL for statistical analysis… Hurray!

SSIS Script Task – Upgrade to SSIS 2012

I upgraded a set of SSIS packages from SSIS 2008 R2 to SSIS 2012 and all of them had a smooth upgrade except for one. I had script tasks in almost all the packages which ran succesfully after I upgraded and tweaked them.

There was one package which was giving me trouble and functionally it dealt with moving, loading and deleting files from various folders. I upgraded the script task, updated all references, modified the project paramaters (as I am using the Project Deployment Model) and when I tried to run the package (Debug mode), the following generic error popped up:

Exception has been thrown by the target of an invocation.
at System.RuntimeMethodHandle._InvokeMethodFast(IRuntimeMethodInfo method, Object target, Object[] arguments, SignatureStruct& sig, MethodAttributes methodAttributes, RuntimeType typeOwner)
at System.RuntimeMethodHandle.InvokeMethodFast(IRuntimeMethodInfo method, Object target, Object[] arguments, Signature sig, MethodAttributes methodAttributes, RuntimeType typeOwner)
at System.Reflection.RuntimeMethodInfo.Invoke(Object obj, BindingFlags invokeAttr, Binder binder, Object[] parameters, CultureInfo culture, Boolean skipVisibilityChecks)
at System.Reflection.RuntimeMethodInfo.Invoke(Object obj, BindingFlags invokeAttr, Binder binder, Object[] parameters, CultureInfo culture)
at System.RuntimeType.InvokeMember(String name, BindingFlags bindingFlags, Binder binder, Object target, Object[] providedArgs, ParameterModifier[] modifiers, CultureInfo culture, String[] namedParams)
at Microsoft.SqlServer.Dts.Tasks.ScriptTask.VSTATaskScriptingEngine.ExecuteScript()


This error message was very generic and did not help me in tracking down the root cause. I decided to put a breakpoint into the Main() method of my script and tried debugging the package again. I was surprised when the script task opened up in VSTA and reached my breakpoint (which contradicted the generic error message – Invocation Error). The issue was with folder security to access the files.

Hope this helps fellow ETL developers who might come across similar issues with a generic invocation error.