Casts - Scalar Types | EdgeDB Docs

Explicit Casts

An explicit cast is just using the angle brackets <type>:

Copy

# Cast a string literal into an integer.
SELECT <int64>"42";

Assignment Casts

Assignment casts happen when inserting new objects. Numeric types will often be automatically cast into the specific type corresponding to the property they are assigned to. This is to avoid extra typing when dealing with numeric value using fewer bits:

Copy

# Automatically cast a literal 42 (which is int64
# by default) into an int16 value.
INSERT MyObject {
    int16_val := 42
};

If assignment casting is supported for a given pair of types, explicit casting of those types is also supported.

Implicit Casts

Implicit casts happen automatically whenever the value type doesn’t match the expected type in an expression. This is mostly supported for numeric casts that don’t incur any potential information loss (in form of truncation), so typically from a less precise type, to a more precise one. The int64 to float64 is a notable exception, which can suffer from truncation of significant digits for very large integer values. There are a few scenarios when implicit casts can occur:

Passing arguments that don’t match exactly the types in the function signature:
Copy
db> ...
WITH x := <float32>12.34 SELECT math::ceil(x);
{13}
The function math::ceil() only takes int64, float64, bigint, or decimal as its argument. So the float32 value will be implicitly cast into a float64 in order to match a valid signature.
Using operands that don’t match exactly the types in the operator signature (this works the same way as for functions):
Copy
db>
SELECT 1 + 2.3;
{3.3}
The operator + is defined only for operands of the same type, so in the expression above the int64 value 1 is implicitly cast into a float64 in order to match the other operand and produce a valid signature.
Mixing different numeric types in a set:
Copy
db>
SELECT {1, 2.3, <float32>4.5} IS float64;
{true, true, true}
All elements in a set have to be of the same type, so the values are cast into float64 as that happens to be the common type to which all the set elements can be implicitly cast. This would work out the same way if UNION was used instead:
Copy
db>
SELECT (1 UNION 2.3 UNION <float32>4.5) IS float64;
{true, true, true}