|Oracle9i SQL Reference
Release 2 (9.2)
Part Number A96540-01
SQL Statements: SAVEPOINT to UPDATE, 3 of 8
SELECT statement or subquery to retrieve data from one or more tables, object tables, views, object views, or materialized views.
If the result (or part of the result) of a
For you to select data from a table or materialized view, the table or materialized view must be in your own schema or you must have the
SELECT privilege on the table or materialized view.
For you to select rows from the base tables of a view,
SELECTprivilege on the view, and
SELECTprivilege on the base tables.
TABLE system privilege also allows you to select data from any table or any materialized view or any view's base table.
To issue a flashback query (using the
flashback_clause), either you must have
FLASHBACK object privilege on the objects in the select list, or you must have
TABLE system privilege.
query_name) lets you assign names to subquery blocks. You can then reference the subquery block multiple places in the query by specifying the query name. Oracle optimizes the query by treating the query name as either an inline view or as a temporary table.
You can specify this clause in any top-level
SELECT statement and in most types of subqueries. The query name is visible to all subsequent subqueries (except the subquery that defines the query name itself) and to the main query.
Restrictions on subquery factoring:
subquery_factoring_clauseas a subquery within another
subquery_factoring_clause, but the
FROMsubquery can contain the
Specify a comment that passes instructions to the optimizer on choosing an execution plan for the statement.
"Hints" and Oracle9i Database Performance Tuning Guide and Reference for the syntax and description of hints
UNIQUE if you want Oracle to return only one copy of each set of duplicate rows selected (these two keywords are synonymous). Duplicate rows are those with matching values for each expression in the select list.
Restrictions on DISTINCT and UNIQUE queries:
UNIQUE, the total number of bytes in all select list expressions is limited to the size of a data block minus some overhead. This size is specified by the initialization parameter
select_listcontains LOB columns.
ALL if you want Oracle to return all rows selected, including all copies of duplicates. The default is
Specify the asterisk to select all columns from all tables, views, or materialized views listed in the
If you are selecting from a table (that is, you specify a table in the
select_list lets you specify the columns you want to retrieve from the database.
query_name, specify a name already specified in the
subquery_factoring_clause. You must have specified the
subquery_factoring_clause in order to specify
query_name in the
select_list. If you specify
query_name in the
select_list, then you also must specify
query_name in the
Specify the object name followed by a period and the asterisk to select all columns from the specified table, view, or materialized view. A query that selects rows from two or more tables, views, or materialized views is a join.
You can use the schema qualifier to select from a table, view, or materialized view in a schema other than your own. If you omit
schema, then Oracle assumes the table, view, or materialized view is in your own schema.
Specify an expression representing the information you want to select. A column name in this list can be qualified with
schema only if the table, view, or materialized view containing the column is qualified with
schema in the
FROM clause. If you specify a member method of an object type, then you must follow the method name with parentheses even if the method takes no arguments.
Specify a different name (alias) for the column expression. Oracle will use this alias in the column heading. The
AS keyword is optional. The alias effectively renames the select list item for the duration of the query. The alias can be used in the
order_by_clause, but not other clauses in the query.
Restrictions on the select_list:
group_by_clausein this statement, then this select list can contain only the following types of expressions:
Oracle9i Database Administrator's Guide for information on key-preserved tables
FROM clause lets you specify the objects from which data is selected.
query_table_expression clause to identify a table, view, materialized view, or partition, or to specify a subquery that identifies the objects.
ONLY clause applies only to views. Specify
ONLY if the view in the
FROM clause is a view belonging to a hierarchy and you do not want to include rows from any of its subviews.
flashback_clause to query past data from a table, view, or materialized view. If you specify
expr must evaluate to a number. If you specify
expr must evaluate to a timestamp value. Oracle returns rows as they existed at the specified system change number or time.
This clause implements SQL-driven flashback, which lets you specify a different system change number or timestamp for each object in the select list. You can also implement session-level flashback using the
Restrictions on the flashback_clause:
flashback_clauseto a remote database object. However, you can include remote objects in a join with local objects to which you apply the
SUBPARTITION, specify the partition or subpartition from which you want to retrieve data. The
partition parameter may be the name of the partition within
table from which to retrieve data or a more complicated predicate restricting retrieval to just one partition of the table.
dblink, specify the complete or partial name for a database link to a remote database where the table, view, or materialized view is located. This database need not be an Oracle database.
If you omit
dblink, then Oracle assumes that the table, view, or materialized view is on the local database.
Restrictions on database links:
REFon a remote table.
AnyDataSetfrom remote tables.
view, specify the name of a table, view, or materialized view from which data is selected.
sample_clause lets you instruct Oracle to select from a random sample of rows from the table, rather than from the entire table.
BLOCK instructs Oracle to perform random block sampling instead of random row sampling.
Oracle9i Database Concepts for a discussion of the difference
sample_percent is a number specifying the percentage of the total row or block count to be included in the sample. The value must be in the range .000001 to (but not including) 100.
Restrictions on the sample_clause:
SAMPLEonly in a query that selects from a single table. Joins are not supported. However, you can achieve the same results by using a
SELECTquery to materialize a sample of an underlying table and then rewrite the original query to refer to the newly created table sample. If you wish, you can write additional queries to materialize samples for other tables.
SAMPLE, Oracle automatically uses cost-based optimization. Rule-based optimization is not supported with this clause.
subquery_restriction_clause lets you restrict the subquery in one of the following ways:
ONLY to indicate that the table or view cannot be updated.
OPTION to indicate that Oracle prohibits any changes to the table or view that would produce rows that are not included in the subquery.
Specify the name of the
OPTION constraint. If you omit this identifier, Oracle automatically assigns the constraint a name of the form
n is an integer that makes the constraint name unique within the database.
table_collection_expression lets you inform Oracle that the value of
collection_expression should be treated as a table for purposes of query and DML operations. The
collection_expression can be a subquery, a column, a function, or a collection constructor. Regardless of its form, it must return a collection value (that is, a value whose type is nested table or varray). This process of extracting the elements of a collection is called collection unnesting.
In earlier releases of Oracle, when
collection_expression can reference columns of tables defined to its left in the
FROM clause. This is called left correlation. Left correlation can occur only in
table_collection_expression. Other subqueries cannot contains references to columns defined outside the subquery.
The optional "(+)" lets you specify that
table_collection_expression should return a row with all fields set to
NULL if the collection is null or empty. The "(+)" is valid only if
collection_expression uses left correlation. The result is similar to that of an outer join.
When you use the "(+)" syntax in the
Specify a correlation name (alias) for the table, view, materialized view, or subquery for evaluating the query. Correlation names are most often used in a correlated query. Other references to the table, view, or materialized view throughout the query must refer to this alias.
This alias is required if the
joined_table syntax to identify tables that are part of a join from which to select data.
join_type indicates the kind of join being performed:
INNERto indicate explicitly that an inner join is being performed. This is the default.
RIGHTto indicate a right outer join.
LEFTto indicate a left outer join.
FULLto indicate a full or two-sided outer join. In addition to the inner join, rows from both tables that have not been returned in the result of the inner join will be preserved and extended with nulls.
FULLto explicitly clarify that an outer join is being performed.
JOIN keyword explicitly states that a join is being performed. You can use this syntax to replace the comma-delimited table expressions used in
WHERE clause joins with
FROM clause join syntax.
ON clause to specify a join condition. Doing so lets you specify join conditions separate from any search or filter conditions in the
When you are specifying an equijoin of columns that have the same name in both tables, the
USING column clause indicates the columns to be used. You can use this clause only if the join columns in both tables have the same name. Do not qualify the column name with a table name or table alias.
In an outer join with the
USING clause, the query returns a single column which is a coalesce of the two matching columns in the join. The coalesce functions as follows:
Restriction on USING column: You cannot specify a LOB column or a collection column in the
CROSS keyword indicates that a cross join is being performed. A cross join produces the cross-product of two relations and is essentially the same as the comma-delimited Oracle notation.
NATURAL keyword indicates that a natural join is being performed. A natural join is based on all columns in the two tables that have the same name. It selects rows from the two tables that have equal values in the relevant columns. When specifying columns that are involved in the natural join, do not qualify the column name with a table name or table alias.
On occasion, the table pairings in natural or cross joins may be ambiguous. For example:
can be interpreted in either of the following ways:
a NATURAL LEFT JOIN (b LEFT JOIN c ON b.c1 = c.c1) (a NATURAL LEFT JOIN b) LEFT JOIN c ON b.c1 = c.c1
To avoid this ambiguity, you can use parentheses to specify the pairings of joined tables. In the absence of such parentheses, Oracle uses left associativity, pairing the tables from left to right.
Restriction on natural joins: You cannot specify a LOB column or a collection column as part of a natural join.
WHERE condition lets you restrict the rows selected to those that satisfy one or more conditions. For
condition, specify any valid SQL condition.
Chapter 5, "Conditions" for the syntax description of
If you omit this clause, then Oracle returns all rows from the tables, views, or materialized views in the
If this clause refers to a
See Also: "Selecting from a Partition: Example"
hierarchical_query_clause lets you select rows in a hierarchical order.
SELECT statements that contain hierarchical queries can contain the
LEVEL pseudocolumn in the select list.
LEVEL returns the value 1 for a root node, 2 for a child node of a root node, 3 for a grandchild, and so on. The number of levels returned by a hierarchical query may be limited by available user memory.
If you specify this clause, do not specify either
BY, as they will destroy the hierarchical order of the
BY results. If you want to order rows of siblings of the same parent, use the
Specify a condition that identifies the row(s) to be used as the root(s) of a hierarchical query. Oracle uses as root(s) all rows that satisfy this condition. If you omit this clause, then Oracle uses all rows in the table as root rows. The
WITH condition can contain a subquery, but it cannot contain a scalar subquery expression.
Specify a condition that identifies the relationship between parent rows and child rows of the hierarchy. The
connect_by_condition can be any condition as described in Chapter 5, "Conditions". However, it must use the
PRIOR operator to refer to the parent row.
Restriction on the CONNECT BY clause: The
connect_by_condition cannot contain a regular subquery or a scalar subquery expression.
Notes on hierarchical queries:
If you specify a hierarchical query and also specify the
BY clause, then the
BY clause takes precedence over any ordering specified by the hierarchical query, unless you specify the
SIBLINGS keyword in the
The manner in which Oracle processes a
WHERE clause (if any) in a hierarchical query depends on whether the
WHERE clause contains a join:
WHEREpredicate contains a join, Oracle applies the join predicates before doing the
WHEREclause does not contain a join, Oracle applies all predicates other than the
BYpredicates after doing the
BYprocessing without affecting the other rows of the hierarchy.
BY clause if you want Oracle to group the selected rows based on the value of
expr(s) for each row and return a single row of summary information for each group. If this clause contains
ROLLUP extensions, then Oracle produces superaggregate groupings in addition to the regular groupings.
Expressions in the
BY clause can contain any columns of the tables, views, or materialized views in the
FROM clause, regardless of whether the columns appear in the select list.
BY clause groups rows but does not guarantee the order of the result set. To order the groupings, use the
ROLLUP operation in the
simple_grouping_clause groups the selected rows based on the values of the first n, n-1, n-2, ... 0 expressions in the
BY specification, and returns a single row of summary for each group. You can use the
ROLLUP operation to produce subtotal values by using it with the
SUM function. When used with
ROLLUP generates subtotals from the most detailed level to the grand total. Aggregate functions such as
COUNT can be used to produce other kinds of superaggregates.
For example, given three expressions (n=3) in the
ROLLUP clause of the
simple_grouping_clause, the operation results in n+1 = 3+1 = 4 groupings.
Rows grouped on the values of the first 'n' expressions are called regular rows, and the others are called superaggregate rows.
Oracle9i Data Warehousing Guide for information on using
CUBE operation in the
simple_grouping_clause groups the selected rows based on the values of all possible combinations of expressions in the specification, and returns a single row of summary information for each group. You can use the
CUBE operation to produce cross-tabulation values.
For example, given three expressions (n=3) in the
CUBE clause of the
simple_grouping_clause, the operation results in 2n = 23 = 8 groupings. Rows grouped on the values of 'n' expressions are called regular rows, and the rest are called superaggregate rows.
SETS are a further extension of the
BY clause that let you specify multiple groupings of data. Doing so facilitates efficient aggregation by pruning the aggregates you do not need. You specify just the desired groups, and Oracle does not need to perform the full set of aggregations generated by
ROLLUP. Oracle computes all groupings specified in the
SETS clause and combines the results of individual groupings with a
ALL operation. The
ALL means that the result set can include duplicate rows.
BY clause, you can combine expressions in various ways:
CUBEoperations with commas so that Oracle combines them into a single
BYclause. The result is a cross-product of groupings from each grouping set.
HAVING clause to restrict the groups of returned rows to those groups for which the specified
TRUE. If you omit this clause, then Oracle returns summary rows for all groups.
HAVING after the
hierarchical_query_clause. If you specify both
HAVING, then they can appear in either order.
Restriction on the HAVING clause: The
HAVING condition cannot contain a scalar subquery expression.
group_by_clausereferences any object type columns, then the query will not be parallelized.
These set operators combine the rows returned by two
SELECT statements into a single result. The number and datatypes of the columns selected by each component query must be the same, but the column lengths can be different. The names of the columns in the result set are the names of the expressions in the select list preceding the set operator.
If you combine more than two queries with set operators, then Oracle evaluates adjacent queries from left to right. You can use parentheses to specify a different order of evaluation.
"The UNION [ALL], INTERSECT, MINUS Operators" for information on these operators
Restrictions on set operators:
VARRAY, or nested table.
MINUSoperators are not valid on
for_update_clausewith these set operators.
subqueryof these operators.
BY clause to order rows returned by the statement. Without an
order_by_clause, no guarantee exists that the same query executed more than once will retrieve rows in the same order.
SIBLINGS keyword is valid only if you also specify the
BY preserves any ordering specified in the hierarchical query clause and then applies the
order_by_clause to the siblings of the hierarchy.
expr orders rows based on their value for
expr. The expression is based on columns in the select list or columns in the tables, views, or materialized views in the
position to order rows based on their value for the expression in this position of the select list;
position must be an integer.
"Sorting Query Results" for a discussion of ordering query results
You can specify multiple expressions in the
order_by_clause. Oracle first sorts rows based on their values for the first expression. Rows with the same value for the first expression are then sorted based on their values for the second expression, and so on. Oracle sorts nulls following all others in ascending order and preceding all others in descending order.
Specify whether the ordering sequence is ascending or descending.
ASC is the default.
Specify whether returned rows containing null values should appear first or last in the ordering sequence.
LAST is the default for ascending order, and
FIRST is the default for descending order.
Restrictions on the order_by_clause:
DISTINCToperator in this statement, then this clause cannot refer to columns unless they appear in the select list.
order_by_clausecan contain no more than 255 expressions.
group_by_clausein the same statement, then this
order_by_clauseis restricted to the following expressions:
UPDATE clause lets you lock the selected rows so that other users cannot lock or update the rows until you end your transaction. You can specify this clause only in a top-level
SELECT statement (not in subqueries).
Prior to updating a LOB value, you must lock the row containing the LOB. One way to lock the row is with an embedded
Nested table rows are not locked as a result of locking the parent table rows. If you want the nested table rows to be locked, then you must lock them explicitly.
Restrictions on the for_update_clause:
CURSORoperator, set operators,
group_by_clause, or aggregate functions.
LONGcolumns and sequences referenced in the same statement.
column clause to lock the select rows only for a particular table or view in a join. The columns in the
OF clause only indicate which table or view rows are locked. The specific columns that you specify are not significant. However, you must specify an actual column name, not a column alias. If you omit this clause, then Oracle locks the selected rows from all the tables in the query.
WAIT clauses let you tell Oracle how to proceed if the
SELECT statement attempts to lock a row that is locked by another user.
NOWAIT to return control to you immediately if a lock exists.
WAIT to instruct Oracle to wait
integer seconds for the row to become available, and then return control to you.
If you specify neither
NOWAIT, then Oracle waits until the row is available and then returns the results of the
The following statement creates the query names
avg_cost for the initial query block containing a join, and then uses the query names in the body of the main query.
WITH dept_costs AS ( SELECT department_name, SUM(salary) dept_total FROM employees e, departments d WHERE e.department_id = d.department_id GROUP BY department_name), avg_cost AS ( SELECT SUM(dept_total)/COUNT(*) avg FROM dept_costs) SELECT * FROM dept_costs WHERE dept_total > (SELECT avg FROM avg_cost) ORDER BY department_name; DEPARTMENT_NAME DEPT_TOTAL ------------------------------ ---------- Sales 313800 Shipping 156400
The following statement selects rows from the
employees table with the department number of 30:
The following statement selects the name, job, salary and department number of all employees except purchase clerks from department number 30:
SELECT last_name, job_id, salary department_id FROM employees WHERE NOT (job_id = 'PU_CLERK' AND department_id = 30);
The following statement selects from subqueries in the
FROM clause and gives departments' total employees and salaries as a decimal value of all the departments:
SELECT a.department_id "Department", a.num_emp/b.total_count "%_Employees", a.sal_sum/b.total_sal "%_Salary" FROM (SELECT department_id, COUNT(*) num_emp, SUM(salary) sal_sum FROM employees GROUP BY department_id) a, (SELECT COUNT(*) total_count, SUM(salary) total_sal FROM employees) b;
You can select rows from a single partition of a partitioned table by specifying the keyword
PARTITION in the
FROM clause. This SQL statement assigns an alias for and retrieves rows from the
sales_q2_2000 partition of the sample table
The following example selects rows from the
oe.orders table for orders earlier than a specified date:
The following query estimates the number of orders in the
The following example creates a sampled subset of the sample table
hr.employees table and then joins the resulting sampled table with
departments. This operation circumvents the restriction that you cannot specify the
sample_clause in join queries:
CREATE TABLE sample_emp AS SELECT employee_id, department_id FROM employees SAMPLE(10); SELECT e.employee_id FROM sample_emp e, departments d WHERE e.department_id = d.department_id AND d.department_name = 'Sales';
The following statements show a current value from the sample table
hr.employees and then changes the value:
SELECT salary FROM employees WHERE last_name = 'Chung'; SALARY ---------- 3800 UPDATE employees SET salary = 4000 WHERE last_name = 'Chung'; 1 row updated. SELECT salary FROM employees WHERE last_name = 'Chung'; SALARY ---------- 4000
To learn what the value was before the update, you can use the following flashback query:
SELECT salary FROM employees AS OF TIMESTAMP (SYSTIMESTAMP - INTERVAL '1' DAY) WHERE last_name = 'Chung'; SALARY ---------- 3800
To revert to the earlier value, use the flashback query as the subquery of another
UPDATE employees SET salary = (SELECT salary FROM employees AS OF TIMESTAMP (SYSTIMESTAMP - INTERVAL '1' DAY) WHERE last_name = 'Chung') WHERE last_name = 'Chung'; 1 row updated. SELECT salary FROM employees WHERE last_name = 'Chung'; SALARY ---------- 3800
To return the minimum and maximum salaries for each department in the
employees table, issue the following statement:
To return the minimum and maximum salaries for the clerks in each department, issue the following statement:
SELECT department_id, MIN(salary), MAX (salary) FROM employees WHERE job_id = 'PU_CLERK' GROUP BY department_id;
To return the number of employees and their average yearly salary across all possible combinations of department and job category, issue the following query on the sample tables
SELECT DECODE(GROUPING(department_name), 1, 'All Departments', department_name) AS department_name, DECODE(GROUPING(job_id), 1, 'All Jobs', job_id) AS job_id, COUNT(*) "Total Empl", AVG(salary) * 12 "Average Sal" FROM employees e, departments d WHERE d.department_id = e.department_id GROUP BY CUBE (department_name, job_id); DEPARTMENT_NAME JOB_ID Total Empl Average Sal ------------------------------ ---------- ---------- ----------- Accounting AC_ACCOUNT 1 99600 Accounting AC_MGR 1 144000 Accounting All Jobs 2 121800 Administration AD_ASST 1 52800 . . . All Departments ST_MAN 5 87360 All Departments All Jobs 107 77798.1308
The following example finds the sum of sales aggregated for three precisely specified groups:
(channel_desc, calendar_month_desc, country_id)
SETS syntax, you would have to write less efficient queries with more complicated SQL. For example, you could run three separate queries and
UNION them, or run a query with a
CUBE(channel_desc, calendar_month_desc, country_id) operation and filter out 5 of the 8 groups it would generate.
SELECT channel_desc, calendar_month_desc, co.country_id, TO_CHAR(sum(amount_sold) , '9,999,999,999') SALES$ FROM sales, customers, times, channels, countries co WHERE sales.time_id=times.time_id AND sales.cust_id=customers.cust_id AND sales.channel_id= channels.channel_id AND customers.country_id = co.country_id AND channels.channel_desc IN ('Direct Sales', 'Internet') AND times.calendar_month_desc IN ('2000-09', '2000-10') AND co.country_id IN ('UK', 'US') GROUP BY GROUPING SETS( (channel_desc, calendar_month_desc, co.country_id), (channel_desc, co.country_id), ( calendar_month_desc, co.country_id) ); CHANNEL_DESC CALENDAR CO SALES$ -------------------- -------- -- -------------- Direct Sales 2000-09 UK 1,378,126 Direct Sales 2000-10 UK 1,388,051 Direct Sales 2000-09 US 2,835,557 Direct Sales 2000-10 US 2,908,706 Internet 2000-09 UK 911,739 Internet 2000-10 UK 876,571 Internet 2000-09 US 1,732,240 Internet 2000-10 US 1,893,753 Direct Sales UK 2,766,177 Direct Sales US 5,744,263 Internet UK 1,788,310 Internet US 3,625,993 2000-09 UK 2,289,865 2000-09 US 4,567,797 2000-10 UK 2,264,622 2000-10 US 4,802,459
The following query with a
BY clause defines a hierarchical relationship in which the
employee_id value of the parent row is equal to the
manager_id value of the child row:
In the following
BY clause, the
PRIOR operator applies only to the
employee_id value. To evaluate this condition, Oracle evaluates
employee_id values for the parent row and
commission_pct values for the child row:
SELECT last_name, employee_id, manager_id FROM employees CONNECT BY PRIOR employee_id = manager_id AND salary > commission_pct;
To qualify as a child row, a row must have a
manager_id value equal to the
employee_id value of the parent row and it must have a
salary value greater than its
To return the minimum and maximum salaries for the employees in each department whose lowest salary is less than $5,000, issue the next statement:
SELECT department_id, MIN(salary), MAX (salary) FROM employees GROUP BY department_id HAVING MIN(salary) < 5000; DEPARTMENT_ID MIN(SALARY) MAX(SALARY) ------------- ----------- ----------- 10 4400 4400 30 2500 11000 50 2100 8200 60 4200 9000
To select all salesmen's records from
employees, and order the results by commission in descending order, issue the following statement:
To select information from
employees ordered first by ascending department number and then by descending salary, issue the following statement:
To select the same information as the previous
SELECT and use the positional
BY notation, issue the following statement:
The following statement locks rows in the
employees table with purchasing clerks located in Oxford (
location_id 2500) and locks rows in the
departments table with departments in Oxford that have purchasing clerks:
SELECT e.employee_id, e.salary, e.commission_pct FROM employees e, departments d WHERE job_id = 'SA_REP' AND e.department_id = d.department_id AND location_id = 2500 FOR UPDATE;
The following statement locks only those rows in the
employees table with purchasing clerks located in Oxford (
location_id 2500). No rows are locked in the
SELECT e.employee_id, e.salary, e.commission_pct FROM employees e, departments d WHERE job_id = 'SA_REP' AND e.department_id = d.department_id AND location_id = 2500 FOR UPDATE OF e.salary;
The following statement is legal even though the third value inserted violates the condition of the subquery
INSERT INTO (SELECT department_id, department_name, location_id FROM departments WHERE location_id < 2000) VALUES (9999, 'Entertainment', 2500);
However, the following statement is illegal because it contains the
INSERT INTO (SELECT department_id, department_name, location_id FROM departments WHERE location_id < 2000 WITH CHECK OPTION) VALUES (9999, 'Entertainment', 2500); * ERROR at line 2: ORA-01402: view WITH CHECK OPTION where-clause violation
The following examples show various ways of joining tables in a query. In the first example, an equijoin returns the name and job of each employee and the number and name of the department in which the employee works:
SELECT last_name, job_id, departments.department_id, department_name FROM employees, departments WHERE employees.department_id = departments.department_id; LAST_NAME JOB_ID DEPARTMENT_ID DEPARTMENT_NAME ------------------- ---------- ------------- ---------------------- ... Sciarra FI_ACCOUNT 100 Finance Urman FI_ACCOUNT 100 Finance Popp FI_ACCOUNT 100 Finance ...
You must use a join to return this data because employee names and jobs are stored in a different table than department names. Oracle combines rows of the two tables according to this join condition:
The following equijoin returns the name, job, department number, and department name of all sales managers:
SELECT last_name, job_id, departments.department_id, department_name FROM employees, departments WHERE employees.department_id = departments.department_id AND job_id = 'SA_MAN'; LAST_NAME JOB_ID DEPARTMENT_ID DEPARTMENT_NAME ------------------- ---------- ------------- ----------------------- Russell SA_MAN 80 Sales Partners SA_MAN 80 Sales Errazuriz SA_MAN 80 Sales Cambrault SA_MAN 80 Sales Zlotkey SA_MAN 80 Sales
This query is identical to the preceding example, except that it uses an additional
where_clause condition to return only rows with a
job value of '
To determine who works in the same department as employee '
Lorentz', issue the following statement:
SELECT last_name, department_id FROM employees WHERE department_id = (SELECT department_id FROM employees WHERE last_name = 'Lorentz');
To give all employees in the
employees table a 10% raise if they have changed jobs (that is, if they appear in the
job_history table), issue the following statement:
UPDATE employees SET salary = salary * 1.1 WHERE employee_id IN (SELECT employee_id FROM job_history);
To create a second version of the
new_departments, with only three of the columns of the original table, issue the following statement:
CREATE TABLE new_departments (department_id, department_name, location_id) AS SELECT department_id, department_name, location_id FROM departments;
The following query uses a self join to return the name of each employee along with the name of the employee's manager. (A
WHERE clause is added to shorten the output.)
SELECT e1.last_name||' works for '||e2.last_name "Employees and Their Managers" FROM employees e1, employees e2 WHERE e1.manager_id = e2.employee_id AND e1.last_name LIKE 'R%'; Employees and Their Managers ------------------------------- Rajs works for Mourgos Raphaely works for King Rogers works for Kaufling Russell works for King
The join condition for this query uses the aliases
e2 for the sample table
The following example uses a left outer join to return the names of all departments in the sample schema
hr, even if no employees have been assigned to the departments:
SELECT d.department_id, e.last_name FROM departments d LEFT OUTER JOIN employees e ON d.department_id = e.department_id ORDER BY d.department_id; DEPARTMENT_ID LAST_NAME ------------- ------------------------- 10 Whalen 20 Hartstein 20 Fay 30 Raphaely ... 250 260 270
Users familiar with the traditional Oracle outer joins syntax will recognize the same query in this form:
SELECT d.department_id, e.last_name FROM departments d, employees e WHERE d.department_id = e.department_id(+) ORDER BY d.department_id;
Oracle Corporation strongly recommends that you use the more flexible Oracle9i
FROM clause join syntax shown in the former example.
The left outer join returns all departments, including those without any employees. The same statement with a right outer join returns all employees, including those not yet assigned to a department:
The employee Zeuss was added to the employees table for these examples, and is not part of the sample data.
SELECT d.department_id, e.last_name FROM departments d RIGHT OUTER JOIN employees e ON d.department_id = e.department_id ORDER BY d.department_id; DEPARTMENT_ID LAST_NAME ------------- ------------------------- ... 110 Higgins 110 Gietz Grant Zeuss
It is not clear from this result whether employees Grant and Zeuss have
NULL, or whether their
department_id is not in the
departments table. To determine this requires a full outer join:
SELECT d.department_id as d_dept_id, e.department_id as e_dept_id, e.last_name FROM departments d FULL OUTER JOIN employees e ON d.department_id = e.department_id ORDER BY d.department_id; D_DEPT_ID E_DEPT_ID LAST_NAME ---------- ---------- ------------------------- ... 110 110 Gietz 110 110 Higgins ... 260 270 999 Zeuss Grant
Because the column names in this example are the same in both tables in the join, you can also use the common column feature (the
USING clause) of the join syntax, which coalesces the two matching columns
department_id. The output is the same as for the preceding example:
SELECT department_id AS d_e_dept_id, e.last_name FROM departments d FULL OUTER JOIN employees e USING (department_id) ORDER BY department_id; D_E_DEPT_ID LAST_NAME ----------- ------------------------- ... 110 Higgins 110 Gietz ... 260 270 Grant Zeuss
You can perform DML operations on nested tables only if they are defined as columns of a table. Therefore, when the
query_table_expr_clause of an
UPDATE statement is a
table_collection_expression, the collection expression must be a subquery that selects the table's nested table column. The examples that follow are based on this scenario:
CREATE TYPE ProjectType AS OBJECT( pno NUMBER, pname CHAR(31), budget NUMBER); CREATE TYPE ProjectSet AS TABLE OF ProjectType; CREATE TABLE dept_work (dno NUMBER, dname CHAR(31), projs ProjectSet) NESTED TABLE projs STORE AS ProjectSetTable ((Primary Key(Nested_Table_Id, pno)) ORGANIZATION INDEX COMPRESS 1); INSERT INTO dept_work VALUES (1, 'Engineering', ProjectSet());
This example inserts into the 'Engineering' department's '
projs' nested table:
INSERT INTO TABLE(SELECT d.projs FROM dept_work d WHERE d.dno = 1) VALUES (1, 'Collection Enhancements', 10000);
This example updates the 'Engineering' department's '
projs' nested table:
This example deletes from the 'Engineering' department's '
projs' nested table
Suppose the database contains a table
hr_info with columns
mgr, and a column of nested table type
people which has
sal columns. You could get all the rows from
hr_info and all the rows from
people using the following statement:
Now suppose that
people is not a nested table column of
hr_info, but is instead a separate table with columns
hiredate, and sal. You can extract the same rows as in the preceding example with this statement:
SELECT t1.department, t2.* FROM hr_info t1, TABLE(CAST(MULTISET( SELECT t3.name, t3.dept, t3.sal FROM people t3 WHERE t3.dept = t1.dept) AS NESTED_PEOPLE)) t2;
Finally, suppose that
people is neither a nested table column of table
hr_info nor a table itself. Instead, you have created a function
people_func that extracts from various sources the name, department, and salary of all employees. You can get the same information as in the preceding examples with the following query:
Oracle9i Application Developer's Guide - Fundamentals for more examples of collection unnesting.
The following statement returns all employees in hierarchical order. The root row is defined to be the employee whose job is '
AD_VP'. The child rows of a parent row are defined to be those who have the employee number of the parent row as their manager number.
SELECT LPAD(' ',2*(LEVEL-1)) || last_name org_chart, employee_id, manager_id, job_id FROM employees START WITH job_id = 'AD_VP' CONNECT BY PRIOR employee_id = manager_id; ORG_CHART EMPLOYEE_ID MANAGER_ID JOB_ID ------------------ ----------- ---------- ---------- Kochhar 101 100 AD_VP Greenberg 108 101 FI_MGR Faviet 109 108 FI_ACCOUNT Chen 110 108 FI_ACCOUNT Sciarra 111 108 FI_ACCOUNT Urman 112 108 FI_ACCOUNT Popp 113 108 FI_ACCOUNT Whalen 200 101 AD_ASST Mavris 203 101 HR_REP Baer 204 101 PR_REP Higgins 205 101 AC_MGR Gietz 206 205 AC_ACCOUNT De Haan 102 100 AD_VP Hunold 103 102 IT_PROG Ernst 104 103 IT_PROG Austin 105 103 IT_PROG Pataballa 106 103 IT_PROG Lorentz 107 103 IT_PROG
The following statement is similar to the previous one, except that it does not select employees with the job '
SELECT LPAD(' ',2*(LEVEL-1)) || last_name org_chart, employee_id, manager_id, job_id FROM employees WHERE job_id != 'FI_MGR' START WITH job_id = 'AD_VP' CONNECT BY PRIOR employee_id = manager_id; ORG_CHART EMPLOYEE_ID MANAGER_ID JOB_ID ------------------ ----------- ---------- ---------- Kochhar 101 100 AD_VP Faviet 109 108 FI_ACCOUNT Chen 110 108 FI_ACCOUNT Sciarra 111 108 FI_ACCOUNT Urman 112 108 FI_ACCOUNT Popp 113 108 FI_ACCOUNT Whalen 200 101 AD_ASST Mavris 203 101 HR_REP Baer 204 101 PR_REP Higgins 205 101 AC_MGR Gietz 206 205 AC_ACCOUNT De Haan 102 100 AD_VP Hunold 103 102 IT_PROG Ernst 104 103 IT_PROG Austin 105 103 IT_PROG Pataballa 106 103 IT_PROG Lorentz 107 103 IT_PROG
Oracle does not return the manager
greenberg, although it does return employees who are managed by
The following statement is similar to the first one, except that it uses the
LEVEL pseudocolumn to select only the first two levels of the management hierarchy:
SELECT LPAD(' ',2*(LEVEL-1)) || last_name org_chart, employee_id, manager_id, job_id FROM employees START WITH job_id = 'AD_PRES' CONNECT BY PRIOR employee_id = manager_id AND LEVEL <= 2; ORG_CHART EMPLOYEE_ID MANAGER_ID JOB_ID ------------------ ----------- ---------- ---------- King 100 AD_PRES Kochhar 101 100 AD_VP De Haan 102 100 AD_VP Raphaely 114 100 PU_MAN Weiss 120 100 ST_MAN Fripp 121 100 ST_MAN Kaufling 122 100 ST_MAN Vollman 123 100 ST_MAN Mourgos 124 100 ST_MAN Russell 145 100 SA_MAN Partners 146 100 SA_MAN Errazuriz 147 100 SA_MAN Cambrault 148 100 SA_MAN Zlotkey 149 100 SA_MAN Hartstein 201 100 MK_MAN
This example shows a query that joins the
departments table on the local database with the
employees table on the
SELECT last_name, department_name FROM employees@houston, departments WHERE employees.department_id = departments.department_id;
The following examples show the general syntax of a correlated subquery:
SELECT select_list FROM table1 t_alias1 WHERE expr operator (SELECT column_list FROM table2 t_alias2 WHERE t_alias1.column operator t_alias2.column); UPDATE table1 t_alias1 SET column = (SELECT expr FROM table2 t_alias2 WHERE t_alias1.column = t_alias2.column); DELETE FROM table1 t_alias1 WHERE column operator (SELECT expr FROM table2 t_alias2 WHERE t_alias1.column = t_alias2.column);
The following statement returns data about employees whose salaries exceed their department average. The following statement assigns an alias to
employees, the table containing the salary information, and then uses the alias in a correlated subquery:
SELECT department_id, last_name, salary FROM employees x WHERE salary > (SELECT AVG(salary) FROM employees WHERE x.department_id = department_id) ORDER BY department_id;
For each row of the
employees table, the parent query uses the correlated subquery to compute the average salary for members of the same department. The correlated subquery performs the following steps for each row of the
department_idof the row is determined.
department_idis then used to evaluate the parent query.
The subquery is evaluated once for each row of the
The following statement returns the current date:
You could select
SYSDATE from the
employees table, but Oracle would return 14 rows of the same
SYSDATE, one for every row of the
employees table. Selecting from
DUAL is more convenient.
The following statement increments the employees_seq sequence and returns the new value:
The following statement selects the current value of employees_seq: