CORR()
Correlation coefficient
DOUBLE PRECISION
CORR ( <expr1>, <expr2> )
Parameter | Description |
---|---|
exprN |
Numeric expression.It may contain a table column, a constant, a variable, an expression, a non-aggregate function or a UDF.Aggregate functions are not allowed as expressions. |
The CORR
function return the correlation coefficient for a pair of numerical expressions.
The function CORR(<expr1>, <expr2>)
is equivalent to
COVAR_POP(<expr1>, <expr2>) / (STDDEV_POP(<expr2>) * STDDEV_POP(<expr1>))
This is also known as the Pearson correlation coefficient.
In a statistical sense, correlation is the degree to which a pair of variables are linearly related.A linear relation between variables means that the value of one variable can to a certain extent predict the value of the other.The correlation coefficient represents the degree of correlation as a number ranging from -1 (high inverse correlation) to 1 (high correlation).A value of 0 corresponds to no correlation.
If the group or window is empty, or contains only NULL
values, the result will be NULL
.
CORR
Examplesselect
corr(alength, aheight) AS c_corr
from measure
COVAR_POP()
Population covariance
DOUBLE PRECISION
COVAR_POP ( <expr1>, <expr2> )
Parameter | Description |
---|---|
exprN |
Numeric expression.It may contain a table column, a constant, a variable, an expression, a non-aggregate function or a UDF.Aggregate functions are not allowed as expressions. |
The function COVAR_POP
returns the population covariance for a pair of numerical expressions.
The function COVAR_POP(<expr1>, <expr2>)
is equivalent to
(SUM(<expr1> * <expr2>) - SUM(<expr1>) * SUM(<expr2>) / COUNT(*)) / COUNT(*)
If the group or window is empty, or contains only NULL
values, the result will be NULL
.