Storage Vectors

The second collection type we’ll look at is StorageVec<T>. Just like vectors on the heap (i.e. Vec<T>), storage vectors allow you to store more than one value in a single data structure where each value is assigned an index and can only store values of the same type. However, unlike Vec<T>, the elements of a StorageVec are stored in persistent storage, and consecutive elements are not necessarily stored in storage slots that have consecutive keys.

In order to use StorageVec<T>, you must first import StorageVec as follows:

use std::storage::StorageVec;

Another major difference between Vec<T> and StorageVec<T> is that StorageVec<T> can only be used in a contract because only contracts are allowed to access persistent storage.

Creating a New Storage Vector

To create a new empty storage vector, we have to declare the vector in a storage block as follows:

    v: StorageVec<u64> = StorageVec {
    },

Just like any other storage variable, two things are required when declaring a StorageVec: a type annotation and an initializer. The initializer is just an empty struct of type StorageVec because StorageVec<T> itself is an empty struct! Everything that is interesting about StorageVec<T> is implemented in its methods.

Storage vectors, just like Vec<T>, are implemented using generics which means that the StorageVec<T> type provided by the standard library can hold any type. When we create a storage vector to hold a specific type, we can specify the type within angle brackets. In the example above, we’ve told the Sway compiler that the StorageVec<T> in v will hold elements of the u64 type.

Updating a Storage Vector

To add elements to a storage vector, we can use the push method, as shown below:

    #[storage(read, write)]fn push_to_storage_vec() {
        storage.v.push(5);
        storage.v.push(6);
        storage.v.push(7);
        storage.v.push(8);
    }

Note two details here. First, in order to use push, we need to first access the vector using the storage keyword. Second, because push requires accessing storage, a storage annotation is required on the ABI function that calls push. While it may seem that #[storage(write)] should be enough here, the read annotation is also required because each call to push requires reading (and then updating) the length of the storage vector which is also stored in persistent storage.

Note: The storage annotation is also required for any private function defined in the contract that tries to push into the vector.

Note: There is no need to add the mut keyword when declaring a StorageVec<T>. All storage variables are mutable by default.

Reading Elements of Storage Vectors

To read a value stored in a vector at a particular index, you can use the get method as shown below:

    #[storage(read)]fn read_from_storage_vec() {
        let third = storage.v.get(2);
        match third {
            Option::Some(third) => {
                log(third)
            },
            Option::None => {
                revert(42)
            },
        }
    }

Note three details here. First, we use the index value of 2 to get the third element because vectors are indexed by number, starting at zero. Second, we get the third element by using the get method with the index passed as an argument, which gives us an Option<T>. Third, the ABI function calling get only requires the annotation #[storage(read)] as one might expect because get does not write to storage.

When the get method is passed an index that is outside the vector, it returns None without panicking. This is particularly useful if accessing an element beyond the range of the vector may happen occasionally under normal circumstances. Your code will then have logic to handle having either Some(element) or None. For example, the index could be coming as a contract method argument. If the argument passed is too large, the method get will return a None value, and the contract method may then decide to revert when that happens or return a meaningful error that tells the user how many items are in the current vector and give them another chance to pass a valid value.

Iterating over the Values in a Vector

To access each element in a vector in turn, we would iterate through all of the valid indices using a while loop and the len method as shown below:

    #[storage(read)]fn iterate_over_a_storage_vec() {
        let mut i = 0;
        while i < storage.v.len() {
            log(storage.v.get(i).unwrap());
            i += 1;
        }
    }

Again, this is quite similar to iterating over the elements of a Vec<T> where we use the method len to return the length of the vector. We also call the method unwrap to extract the Option returned by get. We know that unwrap will not fail (i.e. will not cause a revert) because each index i passed to get is known to be smaller than the length of the vector.

Using an Enum to store Multiple Types

Storage vectors, just like Vec<T>, can only store values that are the same type. Similarly to what we did for Vec<T> in the section Using an Enum to store Multiple Types, we can define an enum whose variants will hold the different value types, and all the enum variants will be considered the same type: that of the enum. This is shown below:

enum TableCell {
    Int: u64,
    B256: b256,
    Boolean: bool,
}

Then we can declare a storage vector in a storage block to hold that enum and so, ultimately, holds different types:

    row: StorageVec<TableCell> = StorageVec {
    },

We can now push different enum variants to the storage vector as follows:

    #[storage(read, write)]fn push_to_multiple_types_storage_vec() {
        storage.row.push(TableCell::Int(3));
        storage.row.push(TableCell::B256(0x0101010101010101010101010101010101010101010101010101010101010101));
        storage.row.push(TableCell::Boolean(true));
    }

Now that we’ve discussed some of the most common ways to use storage vectors, be sure to review the API documentation for all the many useful methods defined on StorageVec<T> by the standard library. For now, these can be found in the source code for StorageVec<T>. For example, in addition to push, a pop method removes and returns the last element, a remove method removes and returns the element at some chosen index within the vector, an insert method inserts an element at some chosen index within the vector, etc.

The Sway Programming Language