Hướng dẫn học Microsoft SQL Server 2008 part 27 pps

Nielsen c09.tex V4 - 07/21/2009 12:40pm Page 222 Part II Manipulating Data With Select Kennedy 3 Kennedy 3 Quinn 2 Kemper 1 Nicholson 0 The advantage of the DIFFERENCE() function is that it broadens the search beyond the first letters. The problem with the function is that it wants to calculate the Soundex value for both parameters, which prevents it from taking advantage of pre-stored Soundex values. Data-Type Conversion Functions Converting data from one data type to another data type is often handled automatically by SQL Server. Many of those conversions are implicit, or automatic. Conversions that are explicit require a CAST() or CONVERT() function: ■ CAST(Input as data type): The ANSI standard SQL means of converting from one data type to another. Even when the conversion can be performed implicitly by SQL Server, using the CAST() function forces the desired data type. CAST() is actually programmed slightly differently than a standard function. Rather than separate the two parameters with a comma (as most functions do), the data passed to the CAST() function is followed by the as keyword and the requested output data type: SELECT CAST(’Away’ AS NVARCHAR(5)) AS ‘Tom Hanks’ Result: TOM HANKS AWAY ■ Another example: SELECT CAST(123 AS NVARCHAR(15)) AS Int2String Result: INT2STRING 123 ■ CONVERT(datatype, expression, style): Returns a value converted to a different data type with optional formatting. The first parameter of this non-ANSI SQL function is the desired data type to be applied to the expression: CONVERT (data type, expression[, style]) 222 www.getcoolebook.com Nielsen c09.tex V4 - 07/21/2009 12:40pm Page 223 Data Types, Expressions, and Scalar Functions 9 The style parameter usually refers to the optional date styles listed in Table 9-3. The style is applied to the output during conversion from datetime to a character-based data type, or to the input during conversion from text to datetime. Generally, the one- or two-digit style pro- vides a two-digit year, and its three-digit counterpart provides a four-digit year. For example, style 1 provides 01/01/03, whereas style 101 provides 01/01/2003. The styles marked with an asterisk (*) in Table 9-3 are the exceptions to this rule. SQL Server also provides numeric formatting styles, but numeric formatting is typically the task of the user interface, not the database. TABLE 9-3 Convert Function Date Styles Style Description Format 0/100 ∗ Default mon dd yyyy hh:miAM (or PM) 1/101 USA mm/dd/yy 2/102 ANSI yy.mm.dd 3/103 British/French dd/mm/yy 4/104 German dd.mm.yy 5/105 Italian dd-mm-yy 6/106 – dd mon yy 7/107 – mon dd, yy 8/108 – hh:mm:ss 9 or 109 ∗ Default+milliseconds mon dd yyyy hh:mi:ss:mmmAM (or PM) 10 or 110 USA mm-dd-yy 11 or 111 Japan yy/mm/dd 12 or 112 ISO yymmdd 13 or 113 ∗ Europe default+milliseconds dd mon yyyy hh:mm:ss:mmm (24h) 14 or 114 – hh:mi:ss:mmm (24h) 20 or 120 ∗ ODBC canonical yyyy-mm-dd hh:mi:ss (24h) 21 or 121 ∗ ODBC canonical + milliseconds yyyy-mm-dd hh:mi:ss.mmm (24h) 126 ISO8601 for XML use yyyy-mm-dd Thh:mm:ss:mmm (no spaces) 127 ISO8601 with time zone Z yyyy-mm-ddThh:mi:ss.mmmZ 130 Kuwaiti dd mon yyyy hh:mi:ss:mmmAM (or PM) 131 Kuwaiti dd/mm/yy hh:mi:ss:mmmAM (or PM) ∗ Both styles return dates with centuries. 223 www.getcoolebook.com Nielsen c09.tex V4 - 07/21/2009 12:40pm Page 224 Part II Manipulating Data With Select Best Practice I n a clean client/server design, the server provides the data in an internal format and the client application formats the data as required by the user. Unformatted data is more independent than formatted data and can be used by more applications. The following code demonstrates the CONVERT() function: SELECT CURRENT_TIMESTAMP AS RawDate, CONVERT (NVARCHAR(25), CURRENT_TIMESTAMP, 100) AS Date100, CONVERT (NVARCHAR(25), CURRENT_TIMESTAMP, 1) AS Date1; Result: RawDate Date100 Date1 2009-11-17 10:27:27.413 Nov 17 2001 10:27AM 11/17/01 An additional data-type conversion function provides a fast way to move data between text and numeric: ■ STR(number, length, decimal): Returns a string from a number: SELECT STR(123,6,2) AS [Str]; Result: Str 123.00 Server Environment Information System functions return information about the current environment. This section covers the more com- monly used system functions: ■ DB_NAME(): Returns the name of the current database, as shown in the following example: SELECT CURRENT_TIMESTAMP AS [Date], DB_NAME() AS [Database]; Result: Date Database 2009-11-15 18:38:50.250 CHA2 224 www.getcoolebook.com Nielsen c09.tex V4 - 07/21/2009 12:40pm Page 225 Data Types, Expressions, and Scalar Functions 9 ■ SERVERPROPERTY(): Several useful pieces of information about the server may be deter- mined from this function, including the following: ■ Collation: The collation type ■ Edition: Enterprise, Developer, Standard, and so on ■ EngineEdition: 2 = Standard, 3 = Enterprise, 4 = Express ■ InstanceName: Null if the default instance ■ ProductVersion: The version number of SQL Server ■ ProductLevel: ‘‘RTM’’ for the initial release-to-manufacturing version, ‘‘SPn’’ for service packs (n is the service pack number), ‘‘CTP’’ for Community Technology Preview versions ■ ServerName: The full server and instance name For example, the following code returns SQL Server engine edition and version information for my current instance of SQL Server: SELECT SERVERPROPERTY (’ServerName’) AS ServerName, SERVERPROPERTY (’Edition’) AS Edition, SERVERPROPERTY (’ProductVersion’) AS ‘ProductVersion’, SERVERPROPERTY (’ProductLevel’) AS ProductLevel; Result: ServerName Edition ProductVersion ProductLevel MAUI Developer Edition (64-bit) 10.0.2520.0 SP1 Summary The previous chapter introduced the basic SELECT statement and query flow. This chapter expanded the concept with expressions and calculations that can be inserted in several places within the query, significantly improving its flexibility. In subsequent chapters, you will see how expressions can receive data from subqueries and user-defined functions, further increasing the power of the query. The next chapter continues the progression of adding capability to the query by joining data from multiple data sources. 225 www.getcoolebook.com Nielsen c09.tex V4 - 07/21/2009 12:40pm Page 226 www.getcoolebook.com Nielsen c10.tex V4 - 07/21/2009 12:42pm Page 227 Merging Data with Joins and Unions IN THIS CHAPTER Applying relational algebra Building scalable code with set-based queries Using inner, outer, complex, and  (theta) joins Merging data vertically with unions T he introduction to this book stated that my purpose was to share the fun of developing with SQL Server. This chapter is it. Making data twist and shout, pulling an answer out of data with a creative query, replacing a few hundred lines of languishing row-by-row iterative code with a single blazingly fast, set-based SQL query — it’s all pure fun and covered here. Relational databases, by their very nature, segment data into several narrow, but long, tables. Seldom does looking at a single table provide meaningful data. Therefore, merging data from multiple tables is an important task for SQL developers. The theory behind merging data sets is relational algebra, as defined by E. F. Codd in 1970. Relational algebra consists of eight relational operators: ■ Restrict: Returns the rows that meet a certain criterion ■ Project: Returns selected columns, or calculated data, from a data set ■ Product: Relational multiplication that returns all possible combinations of data between two data sets ■ Union: Relational addition and subtraction that merges two tables vertically by stacking one table above another table and lining up the columns ■ Intersection: Returns the rows common to two data sets ■ Difference: Returns the rows unique to one data set ■ Join: Returns the horizontal merger of two tables, matching up rows basedoncommondata ■ Divide: The inverse of relational multiplication, returns rows in one data set that match every row in a corresponding data set 227 www.getcoolebook.com Nielsen c10.tex V4 - 07/21/2009 12:42pm Page 228 Part II Manipulating Data With Select In addition, as a method of accomplishing relational algebra, SQL has implemented the following: ■ Subqueries: Similar to a join, but more flexible; the results of the subquery are used in place of an expression, list, or data set within an outer query. In the formal language of relational algebra: ■ A table, or data set, is a relation or entity. ■ Arowisatuple. ■ Acolumnisanattribute. However, I use the common terms of table, row, and column throughout this chapter. Relational theory is now thirty-something and has become better defined over the years as database ven- dors compete with extensions, and database theorists further define the problem of representing reality within a data structure. However, E. F. Codd’s original work is still the foundation of relational database design and implementation. To give credit where credit is due, this entire chapter is based on the work of E. F. Codd and C. J. Date. You can find a complete list of recommended resources in the Resources page on www.SQLServerBible.com. Keep in mind that joins work with more than just tables. As listed in Chapter 8, ‘‘Introducing Basic Query Flow,’’ data sources include local SQL Server tables, subqueries/CTEs, views, table-valued user-defined functions, distributed data sources (other database tables), full-text search results, and XQueries. The reason for writing set-based queries is more than just writing elegant code. Set-based queries scale extremely well. My last consulting contract was developing an OLTP system with a few complexities that required 35,000 transactions per second. The system was able to work at that tps rate because the database design enabled set-based queries within stored procedures. So, while this chapter may seem like it just focuses on writing queries, it’s really setting you up to write better stored procedures. Using Joins In relational algebra, a join is the multiplication of two data sets followed by a restriction of the result so that only the intersection of the two data sets is returned. The whole purpose of the join is to hori- zontally merge two data sets and produce a new result set from the combination by matching rows in one data source to rows in the other data source, as illustrated in Figure 10-1. This section explains the various types of joins and how to use them to select data. By merging the data using the join, the rest of the SQL SELECT statement, including the column expressions, aggregate groupings, and WHERE clause conditions, can access any of the columns or rows from the joined tables. These capabilities are the core and power of SQL. 228 www.getcoolebook.com Nielsen c10.tex V4 - 07/21/2009 12:42pm Page 229 Merging Data with Joins and Unions 10 FIGURE 10-1 A join merges rows from one data set with rows from another data set, creating a new set of rows that includes columns from both. In this diagram, the code, 101, is common to Smith and order number 1, and is used to merge the two original rows into a single result row. Name Smith Code 101 Name Smith Code 101 Code 101 Order 1 Order 1 What’s New with Joins and Unions? J oins and unions are at the heart of SQL, so change here occurs slowly. The only item to watch for with joins and unions is the ANSI 89 style outer joins. If you’re upgrading from SQL Server 2000 directly to SQL Server 2008, you should be warned that ANSI 89 style outer joins (*=, =*) were removed from SQL Server with version 2005. ANSI 89 style inner joins may be a legitimate syntax, but I still don’t recommend using them. I apologize if this sounds too much like your teenager’s math homework, but joins are based on the idea of intersecting data sets. As Figure 10-2 illustrates, a relational join deals with two sets of data that have common values, and it’s these common values that define how the tables intersect. FIGURE 10-2 Relational joins are based on the overlap, or common intersection, of two data sets. Data Set A Data Set B Common Intersection These set diagrams are a type of Venn diagram. For more information about Venn set diagrams, visit www.combinatorics.org/Surveys/ds5/VennEJC.html or Wikipedia. 229 www.getcoolebook.com Nielsen c10.tex V4 - 07/21/2009 12:42pm Page 230 Part II Manipulating Data With Select The intersection simply represents the fact that some common column can connect a row from the first data set to data in the second data set. The common values are typically a primary key and a foreign key, such as these examples from the OBXKites sample database: ■ ContactID between the Contact and [Order] tables ■ OrderID between the [Order] and OrderDetail tables ■ ProductID between the Product and OrderDetail tables SQL includes many types of joins that determine how the rows are selected from the different sides of the intersection. Table 10-1 lists the join types (each is explained in more detail later in this section). TABLE 10-1 Join Types Join Type Query Designer Symbol Definition Inner join Includes only matching rows Left outer join Includes all rows from the left table regardless of whether a match exists, and matching rows from the right table Right outer join Includes all the rows from the right table regardless of whether a match exists, and matching rows from the left table Full outer join Includes all the rows from both tables regardless of whether a match exists  (theta) join Matches rows using a non-equal condition — the symbol shows the actual theta condition (<,>,<=,>=,<>) Cross join No join connection Produces a Cartesian product — a match between each row in data source one with each row from data source two without any conditions or restrictions Inner Joins The inner join is by far the most common join. In fact, it’s also referred to as a common join,andwas originally called a natural join by E. F. Codd. The inner join returns only those rows that represent a match between the two data sets. An inner join is well named because it extracts only data from the inner portion of the intersection of the two overlapping data sets, as illustrated in Figure 10-3. 230 www.getcoolebook.com Nielsen c10.tex V4 - 07/21/2009 12:42pm Page 231 Merging Data with Joins and Unions 10 FIGURE 10-3 The inner join includes only those rows from each side of the join that are contained within the intersection of the two data sources. Inner Join Data Set A Data Set B Common Intersection Building inner joins with the Query Designer Inner joins are easily constructed within Management Studio using the Query Designer UI, as shown in Figure 10-4. Once both tables have been placed in the Diagram pane either using the Add Table func- tion or by dragging the tables from the table list, the join automatically creates the required common joins based on common fields. Any unwanted joins can be removed by selecting the join and pressing Delete. To create a new join between two tables, drag the join column from the first table to the second table. The type of join can be changed by right-clicking on the join type symbol. The Query Designer uses a different symbol for each type of join. The symbol for an inner join, the join diamond, is an accurate illustration of that type of join. Creating inner joins within SQL code Using T- SQL code, joins are specified within the FROM portion of the SELECT statement. The keyword JOIN identifies the second table, and the ON clause defines the common ground between the two tables. The default type of join is an inner join, so the keyword INNER is optional. For clarity, however, I rec- ommend always including it: SELECT * FROM Table1 [INNER] JOIN Table2 ON Table1.column = Table2.column; ON the WEBSITE ON the WEBSITE The sample databases and code from this chapter may be downloaded from www.SQLServerBible.com. 231 www.getcoolebook.com . versions ■ ServerName: The full server and instance name For example, the following code returns SQL Server engine edition and version information for my current instance of SQL Server: SELECT SERVERPROPERTY. Server: SELECT SERVERPROPERTY (’ServerName’) AS ServerName, SERVERPROPERTY (’Edition’) AS Edition, SERVERPROPERTY (’ProductVersion’) AS ‘ProductVersion’, SERVERPROPERTY (’ProductLevel’) AS ProductLevel; Result: ServerName. heart of SQL, so change here occurs slowly. The only item to watch for with joins and unions is the ANSI 89 style outer joins. If you’re upgrading from SQL Server 2000 directly to SQL Server 2008,