RATIO_TO_REPORT - Analytic Functions

In Oracle PL/SQL, RATIO_TO_REPORT is an analytic function which returns the proportion of a value over the total set of values. A statement "RATIO_TO_REPORT of 2 over (1,2,3,4,5)" is (2/15) i.e. 0.133. Note that it returns NULL for NULL values of a column.

Syntax:

RATIO_TO_REPORT(expr) OVER ([ query_partition_clause ])

 It computes the ratio of a value to the sum of a set of values. If expr evaluates to null, then the ratio-to-report value also evaluates to null.

The set of values is determined by the query_partition_clause. If you omit that clause, then the ratio-to-report is computed over all rows returned by the query.

You cannot use RATIO_TO_REPORT or any other analytic function for expr. That is, you cannot nest analytic functions, but you can use other built-in function expressions for expr.




The SQL query below calculates the ratio of an employee's salary over the sum of salaries in his department.

SELECT DEPT, EMPNO, SAL, RATIO_TO_REPORT(SAL) OVER (PARTITION BY DEPT) RATIO
FROM EMPLOYEE

      DEPT      EMPNO        SAL      RATIO
---------- ---------- ---------- ----------
        10        100       2300 .479166667
        10        110       2500 .520833333
        20        120       5400 .382978723
        20        140       3400 .241134752
        20        170       5300 .375886525
        30        180       7300 .776595745
        30        130       2100 .223404255
        40        150       6400          1
        50        160       3200          1

9 rows selected.

SELECT last_name, salary, RATIO_TO_REPORT(salary) OVER () AS rr
   FROM employees
   WHERE job_id = 'PU_CLERK';

LAST_NAME                     SALARY         RR
------------------------- ---------- ----------
Khoo                            3100 .223021583
Baida                           2900 .208633094
Tobias                          2800 .201438849
Himuro                          2600  .18705036
Colmenares                      2500 .179856115

Analytic Function - LAG, LEAD, LISTAGG, COLLECT

Both LAG and LEAD functions have the same usage, as shown below.

LAG  (value_expression [,offset] [,default]) OVER ([query_partition_clause] order_by_clause)
LEAD (value_expression [,offset] [,default]) OVER ([query_partition_clause] order_by_clause)

• value_expression - Can be a column or a built-in function, except for other analytic functions.
• offset - The number of rows preceeding/following the current row, from which the data is to be retrieved. The default value is 1.
• default - The value returned if the offset is outside the scope of the window. The default value is NULL.

Looking at the EMP table, we query the data in salary (SAL) order.

SELECT empno,
       ename,
       job,
       sal
FROM   emp
ORDER BY sal;

     EMPNO ENAME      JOB              SAL
---------- ---------- --------- ----------
      7369 SMITH      CLERK            800
      7900 JAMES      CLERK            950
      7876 ADAMS      CLERK           1100
      7521 WARD       SALESMAN        1250
      7654 MARTIN     SALESMAN        1250
      7934 MILLER     CLERK           1300
      7844 TURNER     SALESMAN        1500
      7499 ALLEN      SALESMAN        1600
      7782 CLARK      MANAGER         2450
      7698 BLAKE      MANAGER         2850
      7566 JONES      MANAGER         2975
      7788 SCOTT      ANALYST         3000
      7902 FORD       ANALYST         3000
      7839 KING       PRESIDENT       5000

SQL>

LAG

The LAG function is used to access data from a previous row. The following query returns the salary from the previous row to calculate the difference between the salary of the current row and that of the previous row. Notice that the ORDER BY of the LAG function is used to order the data by salary.

SELECT empno,
       ename,
       job,
       sal,
       LAG(sal, 1, 0) OVER (ORDER BY sal) AS sal_prev,
       sal - LAG(sal, 1, 0) OVER (ORDER BY sal) AS sal_diff
FROM   emp;

     EMPNO ENAME      JOB              SAL   SAL_PREV   SAL_DIFF
---------- ---------- --------- ---------- ---------- ----------
      7369 SMITH      CLERK            800          0        800
      7900 JAMES      CLERK            950        800        150
      7876 ADAMS      CLERK           1100        950        150
      7521 WARD       SALESMAN        1250       1100        150
      7654 MARTIN     SALESMAN        1250       1250          0
      7934 MILLER     CLERK           1300       1250         50
      7844 TURNER     SALESMAN        1500       1300        200
      7499 ALLEN      SALESMAN        1600       1500        100
      7782 CLARK      MANAGER         2450       1600        850
      7698 BLAKE      MANAGER         2850       2450        400
      7566 JONES      MANAGER         2975       2850        125
      7788 SCOTT      ANALYST         3000       2975         25
      7902 FORD       ANALYST         3000       3000          0
      7839 KING       PRESIDENT       5000       3000       2000

SQL>

In LAG(sal, 1, 0), Second parameter '1' represents how many rows back. When there is no row left it will show default value '0' which is third parameter to this function.

LEAD

The LEAD function is used to return data from the next row. The following query returns the salary from the next row to calulate the difference between the salary of the current row and the following row.

SELECT empno,
       ename,
       job,
       sal,
       LEAD(sal, 1, 0) OVER (ORDER BY sal) AS sal_next,
       LEAD(sal, 1, 0) OVER (ORDER BY sal) - sal AS sal_diff
FROM   emp;

     EMPNO ENAME      JOB              SAL   SAL_NEXT   SAL_DIFF
---------- ---------- --------- ---------- ---------- ----------
      7369 SMITH      CLERK            800        950        150
      7900 JAMES      CLERK            950       1100        150
      7876 ADAMS      CLERK           1100       1250        150
      7521 WARD       SALESMAN        1250       1250          0
      7654 MARTIN     SALESMAN        1250       1300         50
      7934 MILLER     CLERK           1300       1500        200
      7844 TURNER     SALESMAN        1500       1600        100
      7499 ALLEN      SALESMAN        1600       2450        850
      7782 CLARK      MANAGER         2450       2850        400
      7698 BLAKE      MANAGER         2850       2975        125
      7566 JONES      MANAGER         2975       3000         25
      7788 SCOTT      ANALYST         3000       3000          0
      7902 FORD       ANALYST         3000       5000       2000
      7839 KING       PRESIDENT       5000          0      -5000

SQL>

Similar to LAG function, In LEAD(sal, 1, 0), Second parameter '1' represents how many rows forward it should get value. When there is no row left it will show default value '0' which is third parameter to this function.

LISTAGG Analystic Function in 11g Release 2

The LISTAGG analytic function was introduced in Oracle 11g Release 2, making it very easy to aggregate strings. The nice thing about this function is it also allows us to order the elements in the concatenated list. If you are using 11g Release 2 you should use this function for string aggregation.

COLUMN employees FORMAT A50

SELECT deptno, LISTAGG(ename, ',') WITHIN GROUP (ORDER BY ename) AS employees
FROM   emp
GROUP BY deptno;

    DEPTNO EMPLOYEES
---------- --------------------------------------------------
        10 CLARK,KING,MILLER
        20 ADAMS,FORD,JONES,SCOTT,SMITH
        30 ALLEN,BLAKE,JAMES,MARTIN,TURNER,WARD

3 rows selected

WM_CONCAT Built-in Function (Not Supported)

If you are not running 11g Release 2, but are running a version of the database where the WM_CONCAT function is present, then it is a zero effort solution as it performs the aggregation for you. It is actually an example of a user defined aggregate function described below, but Oracle have done all the work for you.

COLUMN employees FORMAT A50

SELECT deptno, wm_concat(ename) AS employees
FROM   emp
GROUP BY deptno;

    DEPTNO EMPLOYEES
---------- --------------------------------------------------
        10 CLARK,KING,MILLER
        20 SMITH,FORD,ADAMS,SCOTT,JONES
        30 ALLEN,BLAKE,MARTIN,TURNER,JAMES,WARD

3 rows selected.

Note. WM_CONCAT is an undocumented function and as such is not supported by Oracle for user applications (MOS Note ID 1336219.1). 

COLLECT function in Oracle 10g

An example on oracle-developer.net uses the COLLECT function in Oracle 10g to get the same result. This method requires a table type and a function to convert the contents of the table type to a string. I've altered his method slightly to bring it in line with this article.

CREATE OR REPLACE TYPE t_varchar2_tab AS TABLE OF VARCHAR2(4000);
/

CREATE OR REPLACE FUNCTION tab_to_string (p_varchar2_tab  IN  t_varchar2_tab,
                                          p_delimiter     IN  VARCHAR2 DEFAULT ',') RETURN VARCHAR2 IS
  l_string     VARCHAR2(32767);
BEGIN
  FOR i IN p_varchar2_tab.FIRST .. p_varchar2_tab.LAST LOOP
    IF i != p_varchar2_tab.FIRST THEN
      l_string := l_string || p_delimiter;
    END IF;
    l_string := l_string || p_varchar2_tab(i);
  END LOOP;
  RETURN l_string;
END tab_to_string;
/

The query below shows the COLLECT function in action.

COLUMN employees FORMAT A50

SELECT deptno,
       tab_to_string(CAST(COLLECT(ename) AS t_varchar2_tab)) AS employees
FROM   emp
GROUP BY deptno;
       
    DEPTNO EMPLOYEES
---------- --------------------------------------------------
        10 CLARK,KING,MILLER
        20 SMITH,JONES,SCOTT,ADAMS,FORD
        30 ALLEN,WARD,MARTIN,BLAKE,TURNER,JAMES
        
3 rows selected.

Source: http://www.oracle-base.com/articles/misc/analytic-functions.php

Analytic Function RANK, DENSE_RANK, FIRST and LAST

This article gives and overview of the RANK, DENSE_RANK, FIRST and LAST analytic functions:-
• RANK
• DENSE_RANK
• FIRST and LAST

RANK:
Let's assume we want to assign a sequential order, or rank, to people within a department based on salary, we might use the RANK function like.

SELECT empno, deptno, sal, 
RANK() OVER (PARTITION BY deptno ORDER BY sal) "rank"
FROM emp;

EMPNO DEPTNO SAL rank
---------- ---------- ---------- ----------
7934 10 1300 1
7782 10 2450 2
7839 10 5000 3
7369 20 800 1
7876 20 1100 2
7566 20 2975 3
7788 20 3000 4
7902 20 3000 4
7900 30 950 1
7654 30 1250 2
7521 30 1250 2
7844 30 1500 4
7499 30 1600 5
7698 30 2850 6

SQL>

What we see here is where two people have the same salary they are assigned the same rank. When multiple rows share the same rank the next rank in the sequence is not consecutive.

DENSE_RANK:
The DENSE_RANK function acts like the RANK function except that it assigns consecutive ranks.

SELECT empno, deptno, sal,
DENSE_RANK() OVER (PARTITION BY deptno ORDER BY sal) "rank"
FROM emp;

EMPNO DEPTNO SAL rank
---------- ---------- ---------- ----------
7934 10 1300 1
7782 10 2450 2
7839 10 5000 3
7369 20  800 1
7876 20 1100 2
7566 20 2975 3
7788 20 3000 4
7902 20 3000 4
7900 30  950 1
7654 30 1250 2
7521 30 1250 2
7844 30 1500 3
7499 30 1600 4
7698 30 2850 5

SQL>

FIRST and LAST:
The FIRST and LAST functions can be used to return the first or last value from an ordered sequence. Say we want to display the salary of each employee, along with the lowest and highest within their department we may use something like.

SELECT empno, deptno, sal,
MIN(sal) KEEP (DENSE_RANK FIRST ORDER BY sal) OVER (PARTITION BY deptno) "Lowest",
MAX(sal) KEEP (DENSE_RANK LAST ORDER BY sal) OVER (PARTITION BY deptno) "Highest"
FROM emp
ORDER BY deptno, sal;

EMPNO DEPTNO SAL Lowest Highest
---------- ---------- ---------- ---------- ----------
7934 10 1300 1300 5000
7782 10 2450 1300 5000
7839 10 5000 1300 5000
7369 20 800 800 3000
7876 20 1100 800 3000
7566 20 2975 800 3000
7788 20 3000 800 3000
7902 20 3000 800 3000
7900 30 950 950 2850
7654 30 1250 950 2850
7521 30 1250 950 2850
7844 30 1500 950 2850
7499 30 1600 950 2850
7698 30 2850 950 2850

FIRST_VALUE and LAST_VALUE Analytic Functions:

This article gives an overview of the FIRST_VALUE and LAST_VALUE analytic functions.
• FIRST_VALUE
• LAST_VALUE

FIRST_VALUE:

The FIRST_VALUE analytic function is similar to the FIRST analytic function, allowing you to return the first result from an ordered set.

SELECT empno, deptno, sal,
FIRST_VALUE(sal) IGNORE NULLS 
OVER (PARTITION BY deptno ORDER BY sal) AS lowest_in_dept
FROM emp;

EMPNO DEPTNO SAL LOWEST_IN_DEPT
---------- ---------- ---------- --------------
7934 10 1300 1300
7782 10 2450 1300
7839 10 5000 1300
7369 20 800 800
7876 20 1100 800
7566 20 2975 800
7788 20 3000 800
7902 20 3000 800
7900 30 950 950
7654 30 1250 950
7521 30 1250 950
7844 30 1500 950
7499 30 1600 950
7698 30 2850 950

SQL>

The "{RESPECT | IGNORE} NULLS" clause indicates if NULLs are considered when determining results.

The windowing clause can be used to alter the window of operation. The following example uses "ROWS 1 PRECEDING" to give a result similar, but not quite the same, to a LAG of 1 row.

SELECT empno, deptno, sal,
FIRST_VALUE(sal) IGNORE NULLS
OVER (PARTITION BY deptno ORDER BY sal ROWS 1 PRECEDING) AS preceding_in_dept
FROM emp;

EMPNO DEPTNO SAL PRECEDING_IN_DEPT
---------- ---------- ---------- -----------------
7934 10 1300 1300
7782 10 2450 1300
7839 10 5000 2450
7369 20 800 800
7876 20 1100 800
7566 20 2975 1100
7788 20 3000 2975
7902 20 3000 3000
7900 30 950 950
7654 30 1250 950
7521 30 1250 1250
7844 30 1500 1250
7499 30 1600 1500
7698 30 2850 1600

SQL>

LAST_VALUE:

The LAST_VALUE analytic function is similar to the LAST analytic function, allowing you to return the last result from an ordered set. Using the default windowing clause the result can be a little unexpected.

SELECT empno, deptno, sal,
LAST_VALUE(sal) IGNORE NULLS
OVER (PARTITION BY deptno ORDER BY sal) AS highest_in_dept
FROM emp;

EMPNO DEPTNO SAL HIGHEST_IN_DEPT
---------- ---------- ---------- ---------------
7934 10 1300 1300
7782 10 2450 2450
7839 10 5000 5000
7369 20 800 800
7876 20 1100 1100
7566 20 2975 2975
7788 20 3000 3000
7902 20 3000 3000
7900 30 950 950
7654 30 1250 1250
7521 30 1250 1250
7844 30 1500 1500
7499 30 1600 1600
7698 30 2850 2850

SQL>

This is because the default windowing clause is "RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW", which in this example means the current row will always be the last value. Altering the windowing clause to "RANGE BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING" gives us the result we probably expected.

SELECT empno, deptno, sal,
LAST_VALUE(sal) IGNORE NULLS
OVER (PARTITION BY deptno ORDER BY sal RANGE BETWEEN
UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING) AS highest_in_dept
FROM emp;

EMPNO DEPTNO SAL HIGHEST_IN_DEPT
---------- ---------- ---------- ---------------
7934 10 1300 5000
7782 10 2450 5000
7839 10 5000 5000
7369 20 800 3000
7876 20 1100 3000
7566 20 2975 3000
7788 20 3000 3000
7902 20 3000 3000
7900 30 950 2850
7654 30 1250 2850
7521 30 1250 2850
7844 30 1500 2850
7499 30 1600 2850
7698 30 2850 2850

SQL>

As with the previous function, the "{RESPECT | IGNORE} NULLS" clause indicates if NULLs are considered when determining results. The default action is RESPECT NULLS.

Analytic Functions Overview - AVG,

Introduced in Oracle 8i, analytic functions, also known as windowing functions, allowed developers to perform tasks in SQL that were previously confined to procedural languages.

• Introduction

• Analytic Function Syntax

o query_partition_clause

o order_by_clause

o windowing_clause

• Using Analytic Functions

Related articles.

• RANK, DENSE_RANK, FIRST and LAST Analytic Functions

• FIRST_VALUE and LAST_VALUE Analytic Functions

• LAG and LEAD Analytic Functions

• LISTAGG Analystic Function in 11g Release 2

• Top-N Queries

Introduction:-

Probably the easiest way to understand analytic functions is to start by looking at aggregate functions. An aggregate function, as the name suggests, aggregates data from several rows into a single result row. For example, we might use the AVG aggregate function to give us an average of all the employee salaries in the EMP table.

SELECT AVG(sal)
FROM emp;

AVG(SAL)
----------
2073.21429

The GROUP BY clause allows us to apply aggregate functions to subsets of rows. For example, we might want to display the average salary for each department.

SELECT deptno, AVG(sal)
FROM emp
GROUP BY deptno
ORDER BY deptno;

DEPTNO AVG(SAL)
---------- ----------
10 2916.66667
20 2175
30 1566.66667

In both cases, the aggregate function reduces the number of rows returned by the query.

Analytic functions also operate on subsets of rows, similar to aggregate functions in GROUP BY queries, but they do not reduce the number of rows returned by the query. For example, the following query reports the salary for each employee, along with the average salary of the employees within the department.

SELECT empno, deptno, sal,
AVG(sal) OVER (PARTITION BY deptno) AS avg_dept_sal
FROM emp;

EMPNO DEPTNO SAL AVG_DEPT_SAL
---------- ---------- ---------- ------------
7782 10 2450 2916.66667
7839 10 5000 2916.66667
7934 10 1300 2916.66667
7566 20 2975 2175
7902 20 3000 2175
7876 20 1100 2175
7369 20 800 2175
7788 20 3000 2175
7521 30 1250 1566.66667
7844 30 1500 1566.66667
7499 30 1600 1566.66667
7900 30 950 1566.66667
7698 30 2850 1566.66667
7654 30 1250 1566.66667

This time AVG is an analytic function, operating on the group of rows defined by the contents of the OVER clause. This group of rows is known as a window, which is why analytic functions are sometimes referred to as window[ing] functions. Notice how the AVG function is still reporting the departmental average, like it did in the GROUP BY query, but the result is present in each row, rather than reducing the total number of rows returned. This is because analytic functions are performed on a result set after all join, WHERE, GROUP BY and HAVING clauses are complete, but before the final ORDER BY operation is performed.

Analytic Function Syntax:

There are some variations in the syntax of the individual analytic functions, but the basic syntax for an analytic function is as follows:

    analytic_function([ arguments ]) OVER (analytic_clause)

The analytic_clause breaks down into the following optional elements.
     [ query_partition_clause ] [ order_by_clause [ windowing_clause ] ]

The sub-elements of the analytic_clause each have their own syntax diagrams, shown here. Rather than repeat the syntax diagrams, the following sections describe what each section of the analytic_clause is used for.

query_partition_clause:

The query_partition_clause divides the result set into partitions, or groups, of data. The operation of the analytic function is restricted to the boundary imposed by these partitions, similar to the way a GROUP BY clause affects the action of an aggregate function. If the query_partition_clause is omitted, the whole result set is treated as a single partition. The following query uses an empty OVER clause, so the average presented is based on all the rows of the result set.

SELECT empno, deptno, sal,
AVG(sal) OVER () AS avg_sal
FROM emp;

EMPNO DEPTNO SAL AVG_SAL
---------- ---------- ---------- ----------
7369 20 800 2073.21429
7499 30 1600 2073.21429
7521 30 1250 2073.21429
7566 20 2975 2073.21429
7654 30 1250 2073.21429
7698 30 2850 2073.21429
7782 10 2450 2073.21429
7788 20 3000 2073.21429
7839 10 5000 2073.21429
7844 30 1500 2073.21429
7876 20 1100 2073.21429
7900 30 950 2073.21429
7902 20 3000 2073.21429
7934 10 1300 2073.21429

SQL>

If we change the OVER clause to include a query_partition_clause based on the department, the averages presented are specifically for the department the employee belongs too.

SELECT empno, deptno, sal,
AVG(sal) OVER (PARTITION BY deptno) AS avg_dept_sal
FROM emp;

EMPNO DEPTNO SAL AVG_DEPT_SAL
---------- ---------- ---------- ------------
7782 10 2450 2916.66667
7839 10 5000 2916.66667
7934 10 1300 2916.66667
7566 20 2975 2175
7902 20 3000 2175
7876 20 1100 2175
7369 20 800 2175
7788 20 3000 2175
7521 30 1250 1566.66667
7844 30 1500 1566.66667
7499 30 1600 1566.66667
7900 30 950 1566.66667
7698 30 2850 1566.66667
7654 30 1250 1566.66667SQL>

order_by_clause:

The order_by_clause is used to order rows, or siblings, within a partition. So if an analytic function is sensitive to the order of the siblings in a partition you should include an order_by_clause. The following query uses the FIRST_VALUE function to return the first salary reported in each department. Notice we have partitioned the result set by the department, but there is no order_by_clause.

SELECT empno, deptno, sal, 

FIRST_VALUE(sal IGNORE NULLS) OVER (PARTITION BY deptno) AS first_sal_in_dept

FROM emp;

EMPNO DEPTNO SAL FIRST_SAL_IN_DEPT
---------- ---------- ---------- -----------------
7782 10 2450 2450
7839 10 5000 2450
7934 10 1300 2450
7566 20 2975 2975
7902 20 3000 2975
7876 20 1100 2975
7369 20  800 2975
7788 20 3000 2975
7521 30 1250 1250
7844 30 1500 1250
7499 30 1600 1250
7900 30 950 1250
7698 30 2850 1250
7654 30 1250 1250

SQL>

Now compare the values of the FIRST_SAL_IN_DEPT column when we include an order_by_clause to order the siblings by ascending salary.

SELECT empno, deptno, sal, 
FIRST_VALUE(sal IGNORE NULLS) OVER (PARTITION BY deptno ORDER BY sal ASC NULLS LAST) AS first_val_in_dept
FROM emp;

EMPNO DEPTNO SAL FIRST_VAL_IN_DEPT
---------- ---------- ---------- -----------------
7934 10 1300 1300
7782 10 2450 1300
7839 10 5000 1300
7369 20 800 800
7876 20 1100 800
7566 20 2975 800
7788 20 3000 800
7902 20 3000 800
7900 30 950 950
7654 30 1250 950
7521 30 1250 950
7844 30 1500 950
7499 30 1600 950
7698 30 2850 950

SQL>

In this case the "ASC NULLS LAST" keywords are unnecessary as ASC is the default for an order_by_clause and NULLS LAST is the default for ASC orders. When ordering by DESC, the default is NULLS FIRST.

It is important to understand how the order_by_clause affects display order. The order_by_clause is guaranteed to affect the order of the rows as they are processed by the analytic function, but it may not always affect the display order. As a result, you must always use a conventional ORDER BY clause in the query if display order is important. Do not rely on any implicit ordering done by the analytic function. Remember, the conventional ORDER BY clause is performed after the analytic processing, so it will always take precedence.

windowing_clause:

We have seen previously the query_partition_clause controls the window, or group of rows, the analytic operates on. The windowing_clause gives some analytic functions a further degree of control over this window within the current partition. The windowing_clause is an extension of the order_by_clause and as such, it can only be used if an order_by_clause is present. 

The windowing_clause has two basic forms:-

RANGE BETWEEN start_point AND end_point
ROWS BETWEEN start_point AND end_point

Possible values for "start_point" and "end_point" are:
• UNBOUNDED PRECEDING: The window starts at the first row of the partition. Only available for start points.
• UNBOUNDED FOLLOWING: The window ends at the last row of the partition. Only available for end points.
• CURRENT ROW :- The window starts or ends at the current row. Can be used as start or end point.
• value_expr PRECEDING :- A physical or logical offset before the current row using a constant or expression that evaluates to a positive numerical value. When used with RANGE, it can also be an interval literal if the order_by_clause uses a DATE column.
• value_expr FOLLOWING :- As above, but an offset after the current row.

The documentation states the start point must always be before the end point, but this is not true, as demonstrated by this rather silly, but valid, query.

SELECT empno, deptno, sal, 
AVG(sal) OVER (PARTITION BY deptno ORDER BY sal ROWS BETWEEN 0 PRECEDING AND 0 PRECEDING) AS avg_of_current_sal
FROM emp;

EMPNO DEPTNO SAL AVG_OF_CURRENT_SAL
---------- ---------- ---------- ------------------
7934 10 1300 1300
7782 10 2450 2450
7839 10 5000 5000
7369 20 800 800
7876 20 1100 1100
7566 20 2975 2975
7788 20 3000 3000
7902 20 3000 3000
7900 30 950 950
7654 30 1250 1250
7521 30 1250 1250
7844 30 1500 1500
7499 30 1600 1600
7698 30 2850 2850

SQL>

In fact, the start point must be before or equal to the end point.

For analytic functions that support the windowing_clause, the default action is "RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW".
The following query is similar to one used previously to report the employee salary and average department salary, but now we have included an order_by_clause so we also get the default windowing_clause. Notice how the average salary is now calculated using only the employees from the same department up to and including the current row.

SELECT empno, deptno, sal, 
AVG(sal) OVER (PARTITION BY deptno ORDER BY sal) AS avg_dept_sal_sofar
FROM emp;

EMPNO DEPTNO SAL AVG_DEPT_SAL_SOFAR
---------- ---------- ---------- ------------------
7934 10 1300 1300
7782 10 2450 1875
7839 10 5000 2916.66667
7369 20 800 800
7876 20 1100 950
7566 20 2975 1625
7788 20 3000 2175
7902 20 3000 2175
7900 30 950 950
7654 30 1250 1150
7521 30 1250 1150
7844 30 1500 1237.5
7499 30 1600 1310
7698 30 2850 1566.66667

SQL>

The following query shows one method for accessing data from previous and following rows within the current row using the windowing_clause. 

This can also be accomplished with LAG and LEAD.

SELECT empno, deptno, sal, 
FIRST_VALUE(sal) OVER (ORDER BY sal ROWS BETWEEN 1 PRECEDING AND CURRENT ROW) AS previous_sal,
LAST_VALUE(sal) OVER (ORDER BY sal ROWS BETWEEN CURRENT ROW AND 1 FOLLOWING) AS next_sal
FROM emp;

EMPNO DEPTNO SAL PREVIOUS_SAL NEXT_SAL
---------- ---------- ---------- ------------ ----------
7369 20 800 800 950
7900 30 950 800 1100
7876 20 1100 950 1250
7521 30 1250 1100 1250
7654 30 1250 1250 1300
7934 10 1300 1250 1500
7844 30 1500 1300 1600
7499 30 1600 1500 2450
7782 10 2450 1600 2850
7698 30 2850 2450 2975
7566 20 2975 2850 3000
7788 20 3000 2975 3000
7902 20 3000 3000 5000
7839 10 5000 3000 5000

SQL>

Using Analytic Functions:
The best way to understand what analytic functions are capable of is to play around with them. List of all analytic functions available in Oracle 11g Release 2. The "*" indicates that these functions allow for the full analytic syntax, including the windowing_clause.

• AVG *
• CORR *
• COUNT *
• COVAR_POP *
• COVAR_SAMP *
• CUME_DIST
• DENSE_RANK
• FIRST
• FIRST_VALUE *
• LAG
• LAST
• LAST_VALUE *
• LEAD
• LISTAGG
• MAX *
• MIN *
• NTH_VALUE *
• NTILE
• PERCENT_RANK
• PERCENTILE_CONT
• PERCENTILE_DISC
• RANK
• RATIO_TO_REPORT
• REGR_ (Linear Regression) Functions *
• ROW_NUMBER
• STDDEV *
• STDDEV_POP *
• STDDEV_SAMP *
• SUM *
• VAR_POP *
• VAR_SAMP *
• VARIANCE *

http://docs.oracle.com/cd/E11882_01/server.112/e26088/functions004.htm