SQL VISUAL QUICKSTART GUIDE- P49 ppt

Listing 15.61 traverses the hierarchy byusing multiple self-joins to trace the chain of command from employee WS3 to the top of the tree.. In Microsoft SQL Server and DB2, to list everyo

Trang 1

Listing 15.61 traverses the hierarchy by

using multiple self-joins to trace the chain

of command from employee WS3 to the top

of the tree See Figure 15.45 for the result.

Unfortunately, you must know the depth of

the hierarchy before you write this query;

use one of the alternatives given in the tip,

if possible

Listing 15.61 Show the full hierarchical relationship of

employee WS3 See Figure 15.45 for the result.

SELECT h1.emp_title || ' < ' ||

h2.emp_title || ' < ' ||

h3.emp_title || ' < ' ||

h4.emp_title

AS chain_of_command FROM hier h1, hier h2, hier h3, hier h4 WHERE h1.emp_title = 'WS3'

AND h1.boss_id = h2.emp_id AND h2.boss_id = h3.emp_id AND h3.boss_id = h4.emp_id;

Listing

chain_of_command -WS3 < DIR1 < VP1 < CEO

Figure 15.45 Result of Listing 15.61.

Trang 2

In Microsoft SQL Server and DB2, to

list everyone who reports to a particular employee (VP1, in this example), either directly or indirectly (through a boss’s boss), use this query:

WITH recurse (emp_title, emp_id) AS (SELECT emp_title,emp_id

FROM hier WHERE emp_title = 'VP1' UNION ALL

SELECT hier.emp_title, hier.emp_id FROM hier, recurse

WHERE recurse.emp_id = hier.boss_id )

SELECT emp_title AS "Works for VP1"

FROM recurse WHERE emp_title <> 'VP1';

In Oracle 10g or later, use the

(nonstan-dard) CONNECT BYsyntax to traverse a hierarchy The following query is equiva-lent to Listing 15.61:

SELECT LTRIM(SYS_CONNECT_BY_PATH(

emp_title, ' < '), ' < ')

AS chain_of_command FROM hier

WHERE LEVEL = 4 START WITH emp_title = 'WS3' CONNECT BY PRIOR boss_id = emp_id;

In Oracle 10g or later, to list everyone

who reports to a particular employee (VP1, in this example), either directly or indirectly (through a boss’s boss), use this query:

SELECT emp_title AS "Works for VP1"

FROM hier WHERE emp_title <> 'VP1' START WITH emp_title = 'VP1' CONNECT BY PRIOR emp_id = boss_id;

✔ Tip

Microsoft Access and

Microsoft SQL Server, change each

||to+ In MySQL, use CONCAT()to

con-catenate strings See “Concatenating

Strings with ||” in Chapter 5

In Microsoft SQL Server and DB2, use

the (standard) recursive WITHclause to

traverse a hierarchy The following query

is equivalent to Listing 15.61 (in SQL

Server, change each ||to+.):

WITH recurse (chain, emp_level,

boss_id) AS

(SELECT

CAST(emp_title

AS VARCHAR(50)),

0,

boss_id

FROM hier

WHERE emp_title = 'WS3'

UNION ALL

SELECT

CAST(recurse.chain || ' < ' ||

hier.emp_title

AS VARCHAR(50)),

recurse.emp_level + 1,

hier.boss_id

FROM hier, recurse

WHERE recurse.boss_id =

hier.emp_id

)

SELECT chain AS chain_of_command

FROM recurse

WHERE emp_level = 3;

Trang 3

Listing 15.62 traverses the hierarchy by

using multiple UNIONs and self-joins to trace

the chain of command for every employee

See Figure 15.46 for the result

Unfortu-nately, you must know the maximum depth

of the hierarchy before you write this query;

use one of the alternatives given in the tip,

if possible

✔ Tip

■ Microsoft Access won’t

run Listing 15.62 because of the restrictions Access puts on join

expressions

To run Listing 15.62 in Microsoft SQL

Server, change each ||to+

To run Listing 15.62 in MySQL, use

CONCAT()instead of ||to concatenate

strings

In Microsoft SQL Server and DB2, use

the (standard) recursive WITHclause to

traverse a hierarchy The following query

is equivalent to Listing 15.62 (in SQL

Server, change each ||to+.):

WITH recurse (emp_title, emp_id) AS

(SELECT

CAST(emp_title

AS VARCHAR(50)), emp_id

FROM hier

WHERE boss_id IS NULL

UNION ALL

SELECT

CAST(recurse.emp_title ||

' > ' ||

h1.emp_title

AS VARCHAR(50)), h1.emp_id

FROM hier h1, recurse

WHERE h1.boss_id =

recurse.emp_id )

SELECT emp_title emp_tree

Listing 15.62 Show the full hierarchal relationship of

every employee See Figure 15.46 for the result.

SELECT chain AS chains_of_command FROM

(SELECT emp_title as chain FROM hier

WHERE boss_id IS NULL UNION

SELECT h1.emp_title || ' > ' ||

h2.emp_title FROM hier h1 INNER JOIN hier h2

ON (h1.emp_id = h2.boss_id) WHERE h1.boss_id IS NULL UNION

h2.emp_title || ' > ' ||

ON (h1.emp_id = h2.boss_id) LEFT OUTER JOIN hier h3

ON (h2.emp_id = h3.boss_id) WHERE h1.emp_title = 'CEO' UNION

h2.emp_title || ' > ' ||

h3.emp_title || ' > ' ||

ON (h1.emp_id = h2.boss_id) INNER JOIN hier h3

ON (h2.emp_id = h3.boss_id) LEFT OUTER JOIN hier h4

ON (h3.emp_id = h4.boss_id) WHERE h1.emp_title = 'CEO' ) chains

WHERE chain IS NOT NULL ORDER BY chain;

Listing

Trang 4

(nonstan-dard) CONNECT BYsyntax to traverse a hierarchy The following query is equiva-lent to Listing 15.62:

SELECT ltrim(SYS_CONNECT_BY_PATH(

emp_title, ' > '),' > ')

AS chains_of_command FROM hier

START WITH boss_id IS NULL CONNECT BY PRIOR emp_id = boss_id;

chains_of_command

-CEO

CEO > VP1

CEO > VP1 > DIR1

CEO > VP1 > DIR1 > WS1

CEO > VP1 > DIR2

CEO > VP2

CEO > VP2 > DIR3

Trang 5

Listing 15.63 uses scalar subqueries to

determine whether each node in the

hierar-chy is a root, branch, or leaf node See

Figure 15.47 for the result A zero in the

result denotes True; nonzero, False

✔ Tip

Microsoft Access, change

eachSIGNtoSGN

(nonstan-dard) CONNECT BYsyntax to traverse a

hierarchy The following query is

equiva-lent to Listing 15.63:

SELECT

emp_title,

(CASE CONNECT_BY_ROOT(emp_title)

WHEN emp_title THEN 1

ELSE 0 END)

AS root_node, (SELECT COUNT(*)

FROM hier h1 WHERE h1.boss_id = hier.emp_id AND hier.boss_id IS NOT NULL AND rownum = 1)

AS branch_node, CONNECT_BY_ISLEAF AS leaf_node

FROM hier

START WITH boss_id IS NULL

CONNECT BY PRIOR emp_id = boss_id

ORDER BY root_node DESC,

branch_node DESC;

Listing 15.63 Determine whether each node is a root,

branch, or leaf node See Figure 15.47 for the result.

SELECT h1.emp_title, (SELECT SIGN(COUNT(*)) FROM hier h2 WHERE h1.emp_id = h2.emp_id AND h2.boss_id IS NULL)

AS root_node, (SELECT SIGN(COUNT(*)) FROM hier h2 WHERE h1.emp_id = h2.boss_id AND h1.boss_id IS NOT NULL)

AS branch_node, (SELECT SIGN(COUNT(*)) FROM hier h2 WHERE 0 = (SELECT COUNT(*) FROM hier h3 WHERE h1.emp_id = h3.boss_id))

AS leaf_node FROM hier h1;

Listing

emp_title root_node branch_node leaf_node

-CEO 1 0 0

VP1 0 1 0

VP2 0 1 0

DIR1 0 1 0

DIR2 0 0 1

DIR3 0 1 0

WS1 0 0 1

WS2 0 0 1

WS3 0 0 1

WS4 0 0 1

WS5 0 0 1

Trang 6

about this book arrow indicating breaks in code, xvi audience for book, xv–xvi

companion website for, xiv knowledge needed by reader, xiv requirements for using book, xviii syntax conventions used, xvi, xvii typographic conventions used, xvi vendor-specific modifications of SQL, xvii absolute pathnames, 3

Access See Microsoft Access

ACID acronym, 403 addition operator (+), 130, 131–132 aggregate functions, 169–192

creating, 171

filtering groups with HAVING, 169, 190–192

inner join combined with GROUP BYand,

215, 217 listing of, 170

returning single values with subqueries, 276

ALLkeyword, 94, 288–290

alternate keys, 39

Symbols

+ (addition operator), 130, 131–132

\ (backslash), 3

[ ] (brackets), 67, 117

/* */ (bracketed comments), 64

^ (caret), 117

/ (division operator), 130, 131–132

\\ (double backslash), 3

= (equal to operator), 101, 110

> (greater than operator), 101

>= (greater than or equal to operator), 101

< (less than operator), 101

<= (less than or equal to operator), 101

* (multiplication operator), 130, 131–132

<> (not equal to operator), 101

( ) (parentheses), 63, 106

% (percent sign) operator, 114

; (semicolon), 62, 63

'(single quotes), xvi, 70, 71

/ (slash), 3

– (subtraction operator), 130, 131–132

_ (underscore) operator, 114

|| (concatenate operator), 134–136

Trang 7

combining with NOTandORoperators, 109

using, 105, 106

ANSI (American National Standards

Institute)

about, xii, xiii

ANSI-89 vs ANSI-92 syntax mode (Access),

5

ANYkeyword

comparing equivalent subqueries using,

301–302

subquery comparisons using, 291–293

approximate numeric types, 75

arguments, 127

arithmetic operators

listing of, 130

order of evaluation of, 133

types of operations using, 127

arrows in code listings, xvi

ASclauses

creating column aliases with, 91–92, 170

table alias creation with, 196–197

ASCclause, 422–428

ASCII encoding, 71

atomic values, 46, 47, 403

authors

author/publisher queries using UNION

operations, 304–307

books written (by publisher), 223–224

books written (by title), 221

calculating greatest number of titles

written

by, 273

comparing values in subqueries with ALL,

289–290

creating table of unpublished, 244, 295

earned royalties by book and, 226–229

finding number of books by, 243, 272

finding pairs by location, 251–252

grouping names of coauthors and sole, 286

join listing cities of publisher and, 238

listed by above-average sales, 278

listing books by, 211

listing by latest date of publications, 272

listing by volume of book sales and, 245

living in different city from publisher, 297,

312

living in publisher’s city, 296, 310 names of sole, 285

outer joins listing all rows with nulls, 239 querying names of co-authors, 284, 285 residing in publisher’s city/state, 213 residing same state as publisher, 276–277 royalty comparisons with subqueries, 279 sorting by genre of writing, 284

sorting by specific location, 249 subquery comparisons using ANY, 292 total book royalties paid by publisher, 233–234

unpublished, 282, 283 using equivalent queries on, 301 writing different genres of books, 298 writing three or more books, 297

adding new rows to and listing, 321–322 creating views of, 388, 389

deleting rows from, 334 structure of, 51, 52

autocommit mode, 404 averages

moving, 407 running, 406

B

backslash (\), 3 balanced trees (B-trees), 382 batch files, 2

batch mode, 2

binary table operations, 36 BLOB (binary large object) data types, 72 books

advances by genre, 216–217 authors writing three or more, 297

by author and listed by publisher, 223–224 calculating running sum and average for, 406

changing prices by genre, 328 comparing values in subqueries with ALL, 289–290

computing running sum of sales, 273

Trang 8

books (continued)

filtering books written by author’s name,

221

finding authors who haven’t written, 282,

283

genres listed by greater sales volume, 250

greatest number written by authors, 273

having above-average sales, 277

listed by sales volume and author, 245

listing authors by latest published, 272

listing by authors, 211

listing number by author, 243, 272

names and IDs of publisher and, 212

place of publication, 214

priced greater than highest price genre, 278

revenues greater than advance, 220

sale prices compared by genre, 280

subqueries listing publishers by genre,

254–255

subquery comparisons using ANY, 292, 293

total royalties for all, 225

types of published by several publishers,

286, 287

updating table values for, 329–330

about, 51

creating sample, 57

listing of books_standard.sqlscript,

57–60

Boolean types, 76

Boyce-Codd normal form (BCNF), 50

bracketed comments (/* */), 64

brackets ([ ])

filter patterns using, 117

using around identifiers, 67

branch nodes, 458

C

calculating statistics

mode, 177

medians, 451–452

overview, 177 running statistics, 406–408, 454–455 sum of set’s values, 174, 179–182 trimmed mean, 432

calendar tables, 414 candidate keys, 39 caret (^), 117

CASEexpression correlated subqueries vs., 274 evaluating conditional values with, 161–164 case sensitivity

changing, 140–141 comparisons and, 140, 173 keyword and identifier, 63 SQL and, xvi

catalogs, 439

character strings See also substrings

case sensitivity of comparisons, 140–141, 173

comparison operators with, 101 concatenating, 134–136 example of string operations, 127 extracting substrings, 137–139 finding length of, 147–148 trimming characters from, 142–146 types of, 70–71

clauses See also specific clauses

about, 62, 63, 64 clients, xv closure property for tables, 36 clustered indexes, 382

Codd, E F., 33, 34, 38 collating sequence, 96 column aliases creating, 91–92 sorting by, 99

columns about, 34, 35, 37 addingUNIQUEconstraints to, 359–362 comparing from similar domains, 199 constraints for, 339–340

creating aliases with ASclauses, 91–92

Trang 9

columns (continued)

defining constraints for, 363–365

derived, 128–129

displaying table definitions for, 316–318

grouping, 171, 183, 184

inserting rows in, 320–322

joining unequal values in, 220

modifying with ALTER TABLE, 373–374

nullability in, 343–345

order in composite indexes, 379

qualifying column names, 194–195, 267

retrieving from SELECTandFROMclauses,

88–90

self-joins within, 247–252

sorting rows by, 95–96

specifying relative position for sorting, 97

subqueries vs joins for working with, 261

two tables on, 211–212

unordered, 35

using simple FOREIGN KEYconstraints,

355–356

when to use indexes for, 378

Command Center (IBM DB2 8.x), 21

Command Editor (IBM DB2 9.x), 22

command-line tools

IBM DB2 db2, 23–26

MySQL mysql, 27–29

Oracle sqlplus, 17–19

PostgreSQLpsql, 30–32

SQL Server osql, 11, 12–13

SQL Server sqlcmd, 10, 15

using SQL with, 2, 3

comments, 62, 64

committing transactions

about, 400

using COMMITstatement, 335, 404

companion website, xiv

comparison operators

ALLmodifications to, 288

ANYmodifications to, 291

listing of, 101

using in subqueries, 275–280

comparisons

case sensitivity of, 140, 173

changing string case, 140–141

composite constraints foreign key, 356–357 primary key, 352 unique, 361–362 composite indexes, 378, 379 concatenate operator (||), 134–136 concurrency transparency, 401 conditions

combining and negating, 105–109, 111–112 equivalent, 112–113

filtering lists with IN, 121–123 filtering with BETWEEN, 118–120 matching row patterns with LIKE, 114–117 re-expressing, 113

types of search, 101 consistency in transactions, 403 constant expressions, 128

constants See literal values

constraints altering or dropping, 375 check, 339, 363–365 column and table, 339–340 foreign key, 339, 353 nullability, 343 primary key, 339, 350 unique, 359–362 using CONSTRAINTclauses, 339–340 converting data types, 157–160

correlated subqueries See also subqueries

comparing author’s royalties with subqueries, 279

computing running sum of book sales, 273

including null values in list of books and authors, 272

listing data in spreadsheet fashion with, 271 qualifying column names in, 272

simple vs., 266 correlation variables, 263

about, 170 comparing equivalent subqueries using, 301–302

creating inner joins with GROUP BYclause and, 215, 230–232

Trang 10

forms of, 178

listing with duplicate rows with, 436

nulls and, 170, 176

statistics using, 177

creating index with, 378–382

unique indexes vs unique constraints, 362

addingUNIQUEconstraints to columns, 360

defined, 337

defining foreign-key constraint in, 353

table creation with, 337, 338, 341–342

366–368

cross joins

accidentally turning inner to, 210

creating, 204–205

defined, 198

curly quotes, xvi

current date and time, 154–155

cursors, 99

D

data control language (DCL) statement, xiii

data definition language (DDL) statement, xiii

data manipulation language (DML)

statement, xiii

data types

approximate numeric, 75

characteristics of, 68–69

comparing subquery values of same, 275

compatibility of for join columns, 199

converting with CAST(), 157–160

datetime, 77–79

exact numeric, 73–74

interval, 80–81

database management systems See DBMSs;

DBMS-specific SQL features

databases See alsobookssample database

command and queries listing tables in, 318

DBMS vs., x

denormalizing, 50 learning to design, 38 picking random rows, 433–434 providing views of data, 386 recovering and restoring data, 401 renaming tables of, 375

rolling back transactions, 400, 403, 404 sample, xviii

dates listing author and publications by latest, 272

sequence tables for incrementing, 413 using in queries, 445

datetime operations data types for, 77–79 example of, 127 extracting part of, 152–153 formatting and, 90 sequence tables for, 412, 413 DB2 CLP windows, 23

db2command-line tool exiting, 26

interactive mode for, 24 script mode for, 25 showing options for, 26 starting, 23

DBMS icon, xvii

DBMSs (database management systems) See also DBMS-specific SQL features

ANSI-89 vs ANSI-92 syntax mode for (Access), 5

command line for, 2 covered by book, xvii databases vs., x determining SQL Server version running, 10 issuing SQL commands to, xii

running in autocommit mode, 404 SQL servers vs desktop, xv support for SQL, ix transactions in, 399–404 using SQL on, 2–3 using with book, xviii working with indexes, 378–382

DBMS-specific SQL features See also specific database programs

Access versions, 5 aggregate functions, 171

Định dạng
Số trang	10
Dung lượng	340,86 KB