Window Functions

This statement is used together with the window function. The OVER clause is used for grouping data and sorting the elements in a group. Window functions are used for generating sequence numbers for the values in the group.

ORDER BY in a window function must be followed by a column name. If it is followed by a number, the number is processed as a constant value and the target column is not ranked.

RANK()

Description: The RANK function is used for generating non-consecutive sequence numbers for the values in each group. The same values have the same sequence number.

Return type: bigint

Example:

      
           gaussdb=# CREATE TABLE rank_t1(a int, b int);

gaussdb=# INSERT INTO rank_t1 VALUES(1,1),(1,1),(1, 2),(1, 3),(2, 4),(2, 5),(3,6);

gaussdb=# SELECT a,b,RANK() OVER(PARTITION BY a ORDER BY b) FROM rank_t1;
 a | b | rank 
---+---+------
 1 | 1 |    1
 1 | 1 |    1
 1 | 2 |    3
 1 | 3 |    4
 2 | 4 |    1
 2 | 5 |    2
 3 | 6 |    1
(7 rows)

gaussdb=# DROP TABLE rank_t1;

ROW_NUMBER()

Description: The ROW_NUMBER function is used for generating consecutive sequence numbers for the values in each group. The same values have different sequence numbers.

Return type: bigint

Example:

      
           gaussdb=# CREATE TABLE row_number_t1(a int, b int);

gaussdb=# INSERT INTO row_number_t1 VALUES(1,1),(1,1),(1, 2),(1, 3),(2, 4),(2, 5),(3,6);

gaussdb=# SELECT a,b,ROW_NUMBER() OVER(PARTITION BY a ORDER BY b) FROM row_number_t1;
 a | b | row_number 
---+---+------------
 1 | 1 |          1
 1 | 1 |          2
 1 | 2 |          3
 1 | 3 |          4
 2 | 4 |          1
 2 | 5 |          2
 3 | 6 |          1
(7 rows)

gaussdb=# DROP TABLE row_number_t1;

DENSE_RANK()

Description: The DENSE_RANK function is used for generating consecutive sequence numbers for the values in each group. The same values have the same sequence number.

Return type: bigint

Example:

      
           gaussdb=# CREATE TABLE dense_rank_t1(a int, b int);

gaussdb=# INSERT INTO dense_rank_t1 VALUES(1,1),(1,1),(1, 2),(1, 3),(2, 4),(2, 5),(3,6);

gaussdb=# SELECT a,b,DENSE_RANK() OVER(PARTITION BY a ORDER BY b) FROM dense_rank_t1;
 a | b | dense_rank 
---+---+------------
 1 | 1 |          1
 1 | 1 |          1
 1 | 2 |          2
 1 | 3 |          3
 2 | 4 |          1
 2 | 5 |          2
 3 | 6 |          1
(7 rows)

gaussdb=# DROP TABLE dense_rank_t1;

PERCENT_RANK()

Description: The PERCENT_RANK function is used for generating corresponding sequence numbers for the values in each group. That is, the function calculates the value according to the formula: Sequence number = (rank - 1) / (totalrows - 1). rank is the corresponding sequence number generated based on the RANK function for the value and totalrows is the total number of elements in a group.

Return type: double precision

Example:

      
           gaussdb=# CREATE TABLE percent_rank_t1(a int, b int);

gaussdb=# INSERT INTO percent_rank_t1 VALUES(1,1),(1,1),(1, 2),(1, 3),(2, 4),(2, 5),(3,6);

gaussdb=# SELECT a,b,PERCENT_RANK() OVER(PARTITION BY a ORDER BY b) FROM percent_rank_t1;
 a | b |   percent_rank   
---+---+------------------
 1 | 1 |                0
 1 | 1 |                0
 1 | 2 | .666666666666667
 1 | 3 |                1
 2 | 4 |                0
 2 | 5 |                1
 3 | 6 |                0
(7 rows)

gaussdb=# DROP TABLE percent_rank_t1;

CUME_DIST()

Description: The CUME_DIST function is used for generating accumulative distribution sequence numbers for the values in each group. That is, the function calculates the value according to the following formula: Sequence number = Number of rows preceding or peer with current row/Total rows.

Return type: double precision

Example:

      
           gaussdb=# CREATE TABLE cume_dist_t1(a int, b int);

gaussdb=# INSERT INTO cume_dist_t1 VALUES(1,1),(1,1),(1, 2),(1, 3),(2, 4),(2, 5),(3,6);

gaussdb=# SELECT a,b,CUME_DIST() OVER(PARTITION BY a ORDER BY b) FROM cume_dist_t1;
 a | b | cume_dist 
---+---+-----------
 1 | 1 |        .5
 1 | 1 |        .5
 1 | 2 |       .75
 1 | 3 |         1
 2 | 4 |        .5
 2 | 5 |         1
 3 | 6 |         1
(7 rows)

gaussdb=# DROP TABLE cume_dist_t1;

NTILE(num_buckets integer)

Description: The NTILE function is used for equally allocating sequential data sets to the buckets whose quantity is specified by num_buckets according to num_buckets integer and allocating the bucket number to each row. Divide the partition as evenly as possible.

Return type: integer

Example:

      
           gaussdb=# CREATE TABLE ntile_t1(a int, b int);

gaussdb=# INSERT INTO ntile_t1 VALUES(1,1),(1,1),(1, 2),(1, 3),(2, 4),(2, 5),(3,6);

gaussdb=# SELECT a,b,NTILE(2) OVER(PARTITION BY a ORDER BY b) FROM ntile_t1;
 a | b | ntile 
---+---+-------
 1 | 1 |     1
 1 | 1 |     1
 1 | 2 |     2
 1 | 3 |     2
 2 | 4 |     1
 2 | 5 |     2
 3 | 6 |     1
(7 rows)

gaussdb=# DROP TABLE ntile_t1;

LAG(value any [, offset integer [, default any ]])

Description: The LAG function is used for generating lag values for the corresponding values in each group. That is, the value of the row obtained by moving forward the row corresponding to the current value by offset (integer) is the sequence number. If the row does not exist after the moving, the result value is the default value. If omitted, offset defaults to 1 and default to NULL. The type of the default value must be the same as that of the value value.