====== Column to CSV ======
  * scalar valued function that converts given column into comma separated values

==== The first (intuitive approach) ====
  * This approach is very intuitive and straightforward, however pretty slow
<code sql>
SELECT @s = isnull(@s + ', ' + CONVERT (varchar(10),NumericValue), CONVERT (varchar(10),NumericValue))
FROM Measurement
ORDER BY ColumnA;
</code>

==== Second approach ====
<code sql>
SELECT SUBSTRING(
(SELECT ',' + CONVERT (varchar(10),FloatValue)
FROM Measurement
ORDER BY ColumnA
FOR XML PATH('')),2,200000) AS CSV
</code>

==== Experimental Comparison ====
  * Comparison was performed on cca 40000 float values.
  * You can see that second approach is considerably faster.

<code>
-- First approach
SQL Server Execution Times:
   CPU time = 1092 ms,  elapsed time = 1090 ms.
-- Second approach
 SQL Server Execution Times:
   CPU time = 63 ms,  elapsed time = 85 ms.

</code>