All Products
Search
Document Center

MaxCompute:CORR

Last Updated:May 16, 2025

CORR is an extension function provided by MaxCompute V2.0 that calculates the Pearson Correlation Coefficient between two columns of data.

Syntax

double corr(<col1>, <col2>)

Parameters

col1 and col2: Required. Specify the two columns of data used to calculate the Pearson correlation coefficient. Multiple types are supported (DOUBLE, BIGINT, INT, SMALLINT, TINYINT, FLOAT, DECIMAL), and the two columns can have different types.

Return value

A value of the DOUBLE type is returned. If a NULL value exists in a row of either input column, that row is not included in the calculation.

Sample data

This section provides sample source data and examples for you to understand how to use the function. In this topic, a table named mf_math_fun_t is created and data is inserted into the table. Sample statements:

create table if not exists mf_math_fun_t(
     int_data     int,
     bigint_data  bigint,
     double_data  double,
     decimal_data decimal,
     float_data   float,
     string_data  string
    );
insert into mf_math_fun_t values
(null, -10, 0.525, 0.525BD, cast(0.525 as float), '10'),
(-20, null, -0.1, -0.1BD, cast(-0.1 as float), '-10'),
(0, -1, null, 20.45BD, cast(-1 as float), '30'),
(-40, 4, 0.89, null, cast(0.89 as float), '-30'),
(5, -50, -1, -1BD, null, '50'),
(-60, 6, 1.5, 1.5BD, cast(1.5 as float), '-50'),
(-1, -70, -7.5, -7.5BD, cast(-7.5 as float),null ),
(-80, 1, -10.2, -10.2BD, cast(-10.2 as float), '-1' ),
(9, -90, 2.58, 2.58BD, cast(2.58 as float), '0'),
(-100, 10, -5.8, -5.8BD, cast(-5.8 as float), '-90');

Query data from the mf_math_fun_t table. Sample statement:

select * from mf_math_fun_t;
-- The following result is returned: 
+------------+-------------+-------------+--------------+------------+-------------+
| int_data   | bigint_data | double_data | decimal_data | float_data | string_data |
+------------+-------------+-------------+--------------+------------+-------------+
| NULL       | -10         | 0.525       | 0.525        | 0.525      | 10          |
| -20        | NULL        | -0.1        | -0.1         | -0.1       | -10         |
| 0          | -1          | NULL        | 20.45        | -1.0       | 30          |
| -40        | 4           | 0.89        | NULL         | 0.89       | -30         |
| 5          | -50         | -1.0        | -1           | NULL       | 50          |
| -60        | 6           | 1.5         | 1.5          | 1.5        | -50         |
| -1         | -70         | -7.5        | -7.5         | -7.5       | NULL        |
| -80        | 1           | -10.2       | -10.2        | -10.2      | -1          |
| 9          | -90         | 2.58        | 2.58         | 2.58       | 0           |
| -100       | 10          | -5.8        | -5.8         | -5.8       | -90         |
+------------+-------------+-------------+--------------+------------+-------------+

Examples

Based on the sample data, calculate the Pearson coefficient between the double_data and float_data columns. Sample statement:

select corr(double_data,float_data) from mf_math_fun_t;

The return value is 1.0.

Related functions

The CORR function is a mathematical function. For more information about functions for data calculation and data transform, see Mathematical functions.