Precision of numbers

[anakrish]

Apologies if this is already captured somewhere. I could not find the answer to this in the docs.

What is the precision of number data type?
Based on experiments, a numbers seems to have 64 63 bits of precision.

check(n) = ((n+1) - 1) == n

a = check(9223372036854775808) #2^63 => true
b = check(18446744073709551616) # 2^64 => false

[charlieegan3]

Hi @anakrish, OPA is using a 64bit signed integer in these cases. So the maximum value is 2^64-1 rather than 2^64. Hope that helps!