We are currently working on an application (window service), which grabs alerts data from a third party web application and parses each alert and process its status, reasons, and source for different attributes and then creates tickets for support staff.
In this project we were logging all detail activities of parsing alerts into our database, which alert was raised, what was the source and IP Address and so on. Our service was grabbing data from that website after five minutes interval. There were nearly hundreds of alerts on each pass. So our log table was getting heavy each day.
Recently I was working on a procedure which required that I marked some items in different categories, and then based on that marking we have to summarize items in different sections. I have extracted the main idea from this procedure for our example. It is using some Boolean algebra to segregate items into different sections.
Recently we had a little tricky problem to solve. We were displaying a report in which we had to bucket numbers in a range, such that only consecutive numbers should be in that range, if any break is there, then a new range should start. Our first solution did not worked as required. Most difficult part was identifying numbers in a sequence, and placing them in a bucket. We could not create any simple T-SQL queries which could easily sort these things out. Then we thought of first capturing the bucket of each number so that we can easily work it out, and that was not possible without cursors. Lets have a look how we did that.
Recently I saw a puzzle on SQL Server Central, I was intrigued to solve it myself, so here is what I was able to accomplish, and in less than 10 minutes.
DECLARE @i int
SET @i = 0
CREATE TABLE #X (val INT)
WHILE @i < 100
SELECT @i = @i + 1
INSERT INTO #X VALUES (@i)
val % 3,
val % 5,
WHEN val % 3 = 0 AND val % 5 = 0 THEN 'BIZZBUZZ'
WHEN val % 3 = 0 THEN 'BIZZ'
WHEN val % 5 = 0 THEN 'BUZZ'
DROP TABLE #X
Today I had to add a new date column to one of my table, and populate it with some random entries. I found a script from SQL Team Blog about generating random numbers. I modified it a little and used it to populate these dates based on another date I had in my table.
Here is modified code:
-- Create the variables for the random number generation
DECLARE @Upper int;
DECLARE @Lower int-- This will create a random number between 1 and 365
SET @Lower = 1 -- The lowest random number
SET @Upper = 365 -- The highest random number
-- we'll use this temp table to assign a random number to each ID value
CREATE TABLE #temp (ID int NOT NULL, RandNum float NULL)
INSERT #temp (ID) SELECT ID FROM FOO
-- now, assign a new random value to each key value in #temp
DECLARE @id int
DECLARE Randomizer CURSOR
FOR SELECT RandNum FROM #temp
FETCH NEXT FROM Randomizer INTO @id
WHILE @@Fetch_Status != -1
UPDATE #temp SET RandNum = rand()
WHERE CURRENT OF Randomizer
FETCH NEXT FROM Randomizer
SET F.NewDate = dateadd(dd, round((T.RandNum * 100), 0), F.AnyDate)
FOO F INNER JOIN #temp T ON F.ID = T.ID
DROP TABLE #temp
Now say your table does not have an identify column, you don’t want to use a temporary table or alter your existing table, but you still would like to have a record number associated with each record. In this case you could use a self join to return a record number for each row. Here is an example that calculates a RecNum column, and displays the LastName for each record in the Northwind.dbo.Employees table. This example uses count(*) to count the number of records that are greater than or equal LastName in this self join.
SELECT COUNT(*) RecNum,
FROM Northwind.dbo.Employees a
JOIN Northwind.dbo.Employees b
ON a.LastName >= b.LastName
The results from this query looks like this:
I’ve run across several applications where the use of IDENTITY values has caused some confusion, specifically in how the @@IDENTITY function operates given external influences. We’ve all likely been faced with how to grab the newly generated IDENTITY value from a SQL Server database table. And there are a 2 popular techniques that accomplish this with SQL Server:
- the @@IDENTITY function
- the SCOPE_IDENTITY() function
Here how you can convert string date to a datetime value:
DECLARE @DateString char(14)
SET @DateString = '20060703074815'
SELECT CAST(SUBSTRING(@DateString, 1, 4) + '-' +
SUBSTRING(@DateString, 5, 2) + '-' +
SUBSTRING(@DateString, 7, 2) + 'T' +
SUBSTRING(@DateString, 9, 2) + ':' +
SUBSTRING(@DateString, 11, 2) + ':' +
SUBSTRING(@DateString, 13, 2) AS datetime)
If you want to find row count for each table in your database without naming any table then use following script:
exec sp_MSforeachtable 'select count(*) as nr_of_rows, ''?'' as table_name from ?'
When you need to find all procedures and function referencing a particular table then use following:
SELECT o.name, t.TABLE_NAME, c.text
FROM syscomments c JOIN sysobjects o ON c.id = o.id JOIN INFORMATION_SCHEMA.Tables t ON c.text LIKE '%'+t.TABLE_NAME+'%'