Gbase 8A OLAP analysis function cume_ Example of dist

GBase 8a Support cume_dist function , Used to calculate less than or equal to or greater than ( according to order The order of ) Percentage of this value .

grammar
cume_disk() over([partition by ] order by [desc])

explain
among partition Whether to open a window , Otherwise, all values shall be treated uniformly .

order by The order of ,ASC（ Default ) = Less than or equal to ,DESC = Greater than or equal to

Examples
Less than or equal to
The following example is the default ascending order , The first line means ： Less than or equal to value 1 The proportion of rows of is 20%

gbase> select id,cume_dist()over(order by id) cr from t2;
±-----±----+
| id | cr |
±-----±----+
| 1 | 0.2 |
| 2 | 0.4 |
| 3 | 0.6 |
| 4 | 0.8 |
| 5 | 1 |
±-----±----+
5 rows in set (Elapsed: 00:00:00.01)
Greater than or equal to
The order is desc, The first line means ： A value greater than or equal to 5 The proportion of rows of is 20%.

gbase> select id,cume_dist()over(order by id desc) cr from t2;
±-----±----+
| id | cr |
±-----±----+
| 5 | 0.2 |
| 4 | 0.4 |
| 3 | 0.6 |
| 2 | 0.8 |
| 1 | 1 |
±-----±----+
5 rows in set (Elapsed: 00:00:00.02)
With window partition
Every partition The percentages are not calculated separately .

gbase> select * from t4;
±-----±-----+
| id | type |
±-----±-----+
| 1 | A |
| 2 | A |
| 3 | A |
| 1 | B |
| 2 | B |
| 3 | B |
| 4 | B |
±-----±-----+
7 rows in set (Elapsed: 00:00:00.00)

gbase> select type,id,cume_dist()over(partition by type order by id) cr from t4;
±-----±-----±------------------+
| type | id | cr |
±-----±-----±------------------+
| A | 1 | 0.333333333333333 |
| A | 2 | 0.666666666666667 |
| A | 3 | 1 |
| B | 1 | 0.25 |
| B | 2 | 0.5 |
| B | 3 | 0.75 |
| B | 4 | 1 |
±-----±-----±------------------+
7 rows in set (Elapsed: 00:00:00.07)
And Percent_rank Comparison of
percent_rank Is to calculate the relative position , Including the starting point 0, and cume_dist It contains equal to , So it won't show up 0. If the data is only 1 That's ok , that percent_rank For what time 0, and cume_dist by 1（100%）.

gbase> select id,cume_dist()over(order by id) cr,percent_rank()over(order by id) pr from t2;
±-----±----±-----+
| id | cr | pr |
±-----±----±-----+
| 1 | 0.2 | 0 |
| 2 | 0.4 | 0.25 |
| 3 | 0.6 | 0.5 |
| 4 | 0.8 | 0.75 |
| 5 | 1 | 1 |
±-----±----±-----+
5 rows in set (Elapsed: 00:00:00.01)
A line of data

gbase> select id,cume_dist()over(order by id) cr,percent_rank()over(order by id) pr from t5;
±-----±—±—+
| id | cr | pr |
±-----±—±—+
| 1 | 1 | 0 |
±-----±—±—+
1 row in set (Elapsed: 00:00:00.02)