Skip to main content

​
module  pixeltable.functions.math

Pixeltable UDFs for mathematical operations. Example: 
import pixeltable as pxt

t = pxt.get_table(...)
t.select(t.float_col.floor()).collect()

​
udf  abs()

Signature
abs(self: pxt.Float) -> pxt.Float
Return the absolute value of the given number. Equivalent to Python builtins.abs().

​
udf  bitwise_and()

Signature
bitwise_and(self: pxt.Int, other: pxt.Int) -> pxt.Int
Bitwise AND of two integers. Equivalent to Python self & other.

​
udf  bitwise_or()

Signature
bitwise_or(self: pxt.Int, other: pxt.Int) -> pxt.Int
Bitwise OR of two integers. Equivalent to Python self | other.

​
udf  bitwise_xor()

Signature
bitwise_xor(self: pxt.Int, other: pxt.Int) -> pxt.Int
Bitwise XOR of two integers. Equivalent to Python self ^ other.

​
udf  ceil()

Signature
ceil(self: pxt.Float) -> pxt.Float
Return the ceiling of the given number. Equivalent to Python float(math.ceil(self)) if self is finite, or self itself if self is infinite. (This is slightly different from the default behavior of math.ceil(self), which always returns an int and raises an error if self is infinite. The behavior in Pixeltable generalizes the Python operator and is chosen to align with the SQL standard.)

​
udf  floor()

Signature
floor(self: pxt.Float) -> pxt.Float
Return the ceiling of the given number. Equivalent to Python float(math.floor(self)) if self is finite, or self itself if self is infinite. (This is slightly different from the default behavior of math.floor(self), which always returns an int and raises an error if self is infinite. The behavior of Pixeltable generalizes the Python operator and is chosen to align with the SQL standard.)

​
udf  pow()

Signature
pow(self: pxt.Int, other: pxt.Int) -> pxt.Float
Raise self to the power of other. Equivalent to Python self ** other.

​
udf  round()

Signature
round(
    self: pxt.Float,
    digits: pxt.Int | None = None
) -> pxt.Float
Round a number to a given precision in decimal digits. Equivalent to Python builtins.round(self, digits or 0). Note that if digits is not specified, the behavior matches builtins.round(self, 0) rather than builtins.round(self); this ensures that the return type is always float (as in SQL) rather than int.