SQL Server has a built in method for easily pulling one value from an XML. To do this you use the method called .value(). The .value method will run an XQuery against the XML specified in the query. This method is scalar, so it will only return 1 value. You cannot use this to return multiple values.
-- Load the XML data in to a variable to work with.
-- This would typically be passed as a parameter to a stored procedure
In the query above, you can see that you can just pass the xpath and the datatype that you want the output to be. This will pull that value out of the XML and put it in the specified format. You can see that after the xpath, there is a funny syntax . This tells SQL Server to grab the first instance of the AnimalName node. You need to do this because the .value() method only returns one value and will not work with a repeating node.
CROSS APPLY is one of those helpful things in SQL Server that most people don’t think of or may not even know about. In this article I’d like to talk about what the APPLY operator is and how we can use it to simplify our sql statements.
The APPLY operator allows you to join a table to a table-valued function. A table-valued function is a function that will return a table with one or more columns. With the apply operator, you can pass values from the first table in to the table-valued function.
There are only 2 types of APPLY operators: CROSS APPLY – Returns records when a value from both sides of the operator match. Like an INNER JOIN. OUTER APPLY – Returns all rows from the other side of the operator and will return the value or NULL from the table-valued function. This is like an OUTER JOIN.
In the example above, you can see that we join to the GetAnimalHabitat function using the CROSS APPLY. You can imagine that this function does a bunch of logic that is not visible in this query. If the GetAnimalHabitat function had 25 lines of code, you can see how this simplifies the above query dramatically.
The APPLY operator can simplify the code, but could be accomplished by joining to a sub query as well. One difference is that the function in the APPLY operator is being executed for every row in the outer table. If you use the APPLY operator, make sure that you test the speed of your query to make sure that it did not degrade performance.
Replacing text in SQL Server is easy. SQL Server has the REPLACE function to perform this task.
SQL Server REPLACE
The REPLACE function in SQL Server has 3 parameters.
Text to search
Text to find
Text to replace with
SELECTREPLACE('Full text to search in','search','replace')ASReplacedText
The output is: Full text to replace in
In the above example, we are searching “Text to search in” for the word “search” and we are replacing it with the word “replace”. Pretty simple, right? Now let’s take it a step further.
SQL Server Case Sensitive Replace
Above we went over how to do a case insensitive replace on a string. (The REPLACE function actually uses the default collation of the input text that it is searching). To turn it in to a SQL Server case sensitive replace, we just need to add one small thing to the end. We need to change the collation of the text we are searching. Learn more about text collation here.
SELECTREPLACE('Full text to search in'COLLATESQL_Latin1_General_CP1_CS_AS,'SEARCH','replace')ASReplacedText
The output is: Full text to search in
This example shows how the “SEARCH” text was not found because we are changing the input text to search collation to SQL_Latin1_General_CP1_CS_AS (which is case sensitive). The input string has “search” in lower case and we were searching for an upper case “SEARCH”.
SQL Rank Vs Row Number is a common question because the two functions are very similar. SQL Server has the two similar built-in functions to assign numbers to the order of the data in the result set. The SQL Server RANK function and the ROW_NUMBER function both can accomplish this, but they do something slightly different. You can follow these links see details on how to use this functions at RANK function and ROW_NUMBER function.
ROW_NUMBER: Returns a sequential number starting at 1 in the order specified. If there are duplicate records, SQL Server will continue the sequence on the duplicated record… ensuring that the row number sequence is never duplicated.
RANK: Returns a sequential number starting at 1 in the order specified. If there are duplicate records, SQL Server will use the same number for the duplicate records. There will then be a gap in the sequence for the duplicated records.
When it comes to ranking data in SQL Server, as of SQL Server 2005 they have built this in to the database. In 2005 the SQL Server team introduced the RANK function. The RANK function allows you to assign a number to each record in a result set based something in the data.
Let’s say that you wanted to apply a rank based on the weight of animals in your zoo (from heaviest to lightest). Elephant would get a number of 1 (because it’s the heaviest) and a tiny turtle would get the number 100 (because it’s the lightest). The number returned would start at 1 and grow for every record. This is done using this awesome RANK function.
The RANK function also has a partition feature that allows you to group the data. This allows the ranking number to reset for every group. Using our example above, we could extend it to show us the heaviest to lightest animals based on the animal type. So there would be a #1 ranking for the mammal, amphibian, bird, and reptile animal types.
Because the ranking number is incremented in the order specified by the ORDER BY clause, what happens if there are duplicates? In our example above, what if there are 2 animals with the same weight? If this occurs, the same ranking number will be applied to each of the duplicate records. So if there were 2 animals with the same weight, they would get the same exact ranking number.
How To Use RANK
The RANK function has 2 parameters. You must always supply the ORDER BY. The PARTITION BY is optional.
PARTITION BY – This is what you would like SQL Server to group your rankings by. In the above example, if we want the ranking number to reset for every different animal types, then we would specify that here. ORDER BY – The ORDER BY is the order that you want your ranking number to be generated in your result set. This is a required parameter. In the example above, we want it from heaviest to lightest. We would specify the column name here that holds the animal weight.
Here is some sample code and the output to show each of the above scenarios.
In the above example, you can see that we are ranking each record by the weight of the animal from heaviest to lightest. We take this a bit further by grouping these by the AnimalType. This is done by passing the PARTITION BY parameter to the function call.