ZIPF

Syntax

ZIPF(<s> , <N> , <gen>)

Purpose

This function returns an integer that follows the Zipf distribution in the range [0, N), where s is the characteristic exponent of the partition.

• s is the characteristic exponent. A larger s value results in a more skewed sequence. When plotted as a curve, the curve becomes steeper.
• s and N must be scalar values that do not change with row iteration. For example, they can be integer or floating-point constants, scalar functions, or expressions like @v1 or 1+@v3 in PL. The value of s ranges from [1, +∞), and the value of N ranges from [1, 16777215].
• The storage and computational resources consumed by the zipf algorithm are related to the value of N. The space complexity of the algorithm is O(N), and the time complexity for generating each integer is O(logN). Therefore, the value of N is limited to the range [1, 16777215].
• gen is a numerical generation function, typically using the RANDOM() function. If the input value is a constant, the zipf() function returns a fixed value.

Examples

The following example demonstrates how to use the ZIPF() function to return an integer that follows the Zipf distribution.

obclient> SELECT ZIPF(1, 10, RANDOM()) FROM TABLE(GENERATOR(6));
+-----------------------+
|  ZIPF(1,10,RANDOM())  |
+-----------------------+
|                     2 |
|                     0 |
|                     0 |
|                     0 |
|                     3 |
|                     3 |
+-----------------------+
6 rows in set

obclient> SELECT ZIPF(1, 10, 0415) FROM TABLE(GENERATOR(6));
+-------------------+
| ZIPF(1, 10, 0415) |
+-------------------+
|                 1 |
|                 1 |
|                 1 |
|                 1 |
|                 1 |
|                 1 |
+-------------------+
6 rows in set

obclient> SELECT ZIPF(ABS(-1), 23, RANDOM()) FROM DUAL;
+-----------------------------+
| ZIPF(ABS(-1),23,RANDOM()) |
+-----------------------------+
|                           9 |
+-----------------------------+
1 row in set