A BigNumber is an object which safely allows mathematical operations on numbers of any magnitude.
Most operations which need to return a value will return a BigNumber and parameters which accept values will generally accept them.
Many functions and methods in this library take in values which can be non-ambiguously and safely converted to a BigNumber. These values can be specified as:
A HexString or a decimal string, either of which may be negative.
A BytesLike Object, such as an Array or Uint8Array.
An existing BigNumber instance.
The constructor of BigNumber cannot be called directly. Instead, Use the static
Returns an instance of a BigNumber for aBigNumberish.
The BigNumber class is immutable, so no operations can change the value it represents.
Returns a BigNumber with the value of BigNumber + otherValue.
Returns a BigNumber with the value of BigNumber - otherValue.
Returns a BigNumber with the value of BigNumber × otherValue.
Returns a BigNumber with the value of BigNumber ÷ divisor.
Returns a BigNumber with the value of the remainder of BigNumber ÷ divisor.
Returns a BigNumber with the value of BigNumber to the power of exponent.
Returns a BigNumber with the value of BigNumber with bits beyond the bitcount least significant bits set to zero.
Two's Complement is an elegant method used to encode and decode fixed-width signed values while efficiently preserving mathematical operations. Most users will not need to interact with these.
Returns a BigNumber with the value of BigNumber converted from twos-complement with bitwidth.
Returns a BigNumber with the value of BigNumber converted to twos-complement with bitwidth.
Returns true if and only if the value of BigNumber is equal to otherValue.
Returns true if and only if the value of BigNumber < otherValue.
Returns true if and only if the value of BigNumber ≤ otherValue.
Returns true if and only if the value of BigNumber > otherValue.
Returns true if and only if the value of BigNumber ≥ otherValue.
Returns true if and only if the value of BigNumber is zero.
This will throw an error if the value is greater than or equal to Number.MAX_SAFE_INTEGER or less than or equal to Number.MIN_SAFE_INTEGER.
Returns the value of BigNumber as a base-10 string.
Returns the value of BigNumber as a base-16,
Returns true if and only if the object is a BigNumber object.
This section is a for a couple of questions that come up frequently.
The first problem many encounter when dealing with Ethereum is the concept of numbers. Most common currencies are broken down with very little granularity. For example, there are only 100 cents in a single dollar. However, there are 1018 wei in a single ether.
To demonstrate how this may be an issue in your code, consider:
To remedy this, all numbers (which can be large) are stored and manipulated as Big Numbers.
The functions parseEther( etherString ) and formatEther( wei ) can be used to convert between string representations, which are displayed to or entered by the user and Big Number representations which can have mathematical operations handled safely.
Everyone has their own favourite Big Number library, and once someone has chosen one, it becomes part of their identity, like their editor, vi vs emacs. There are over 100 Big Number libraries on npm.
One of the biggest differences between the Ethers BigNumber object and other libraries is that it is immutable, which is very important when dealing with the asynchronous nature of the blockchain.
Capturing the value is not safe in async functions, so immutability protects us from easy to make mistakes, which is not possible on the low-level library's objects which supports myriad in-place operations.
Second, the Ethers BigNumber provides all the functionality required internally and should generally be sufficient for most developers while not exposing some of the more advanced and rare functionality. So it will be easier to swap out the underlying library without impacting consumers.
For example, if BN.js was exposed, someone may use the greatest-common-denominator functions, which would then be functionality the replacing library should also provide to ensure anyone depending on that functionality is not broken.
Therefore it must be included regardless, so we leverage that library rather than adding another Big Number library, which would mean two different libraries offering the same functionality.
This has saved about 85kb (80% of this library size) of library size over other libraries which include separate Big Number libraries for various purposes.
Another comment that comes up frequently is the desire to specify a global user-defined Big Number library, which all functions would return.
This becomes problematic since your code may live along side other libraries or code that use Ethers. In fact, even Ethers uses a lot of the public functions internally.
If you, for example, used a library that used
a.plus(b) instead of
a.add(b), this would break Ethers when it tries to compute fees internally, and other libraries likely have similar logic.