COVAR_SAMP() GROUP BY ...

All functions > GROUP BY > COVAR_SAMP() GROUP BY ...

Returns the sample covariance between two expressions.

Syntax

COVAR_SAMP(expr_y, expr_x)
 [ FILTER (WHERE condition) ]
  [ GROUP BY feature [, feature ...] ]

COVAR_SAMP(expr_y, expr_x)
 [ FILTER (WHERE condition) ]
  [ GROUP BY feature [, feature ...] ]

Notes

• Computes the sample covariance (divides by N-1, Bessel's correction)
• NULL values in either expression are excluded from the calculation
• Returns NULL if fewer than 2 non-NULL pairs exist

Examples

SELECT
    f1 := ZIP(ARRAY[1, 2, 3] AS x, ARRAY[2, 4, 6] AS y).TRANSFORM(SELECT COVAR_SAMP(y, x)).UNWRAP_ONE() -- Sample covariance
;