Table of Contents for Difference between abi.encode, abi.encodePacked, abi.encodeWithSignature and encodeWithSelector

There are a few ways to serialise arguments when calling other smart contracts.

You will often see use of abi.encode(), abi.encodepacked(), abi.encodePacked(), abi.encodeWithSelector().

You can use them like this:

(success, result) = address(someDeployedContract)
        .call(abi.encodeWithSignature("foo(string)", "hello, world"));

You can also ‘manually’ calculate the call argument with something like this:

(success, result) = address(someDeployedContract)
        .call(bytes4(keccak256("foo(string)")), "hello, world");

abi.encode()

Encodes the params according to the ABI specs. Params are padded out to 32 bytes.

abi.encodePacked()

Encodes the params using the minimal amount of space.

There will be no padding for static types, if they’re smaller than 32 bytes. Dynamic types are encoded without the length. Encoding for example a uint8 will use 1 byte. You would use this when saving space, not when calling another contract.

abi.encodeWithSignature()

When you already have a function signature (e.g. bytes4(keccak256("foo(string)"))) then you can use encodeWithSignature() to encode the params

abi.encodeWithSelector()

Very similar to encodeWithSignature.

These two are equivalent

• abi.encodeWithSelector(bytes4(keccak256(signature), ...)
• abi.encodeWithSignature(string signature, ...)

You can get the selector of a function) by adding .selector to it:

// assumes someContract is typed as a contract that has `foo` as a function 
(e.g. `foo(uint256)`)
bytes4 selector = someContract.foo.selector;

Spotted a typo or have a suggestion to make this crypto dev article better? Please let me know!