Skip to main content

Working efficiently with storage items

While Substrate’s architecture offers a developer-friendly abstraction (API) layer for storage, making it simpler to persist data, efficient storage design is crucial for performance, upgradeability, and resource utilization. This guide takes a problem-based approach to common issues and best practices for working with runtime storage items.

By the conclusion of this guide, developers will be equipped with a good understanding of storage management, including but not limited to error handling, data validation through the ensure! macro, and complex conditional storage checks.

To help us measure our progress and improve Substrate in Bits content, please fill out our living feedback form. It will only take 2 minutes of your time. Thank you!

Reproducing errors

Environment and project setup

To follow along with this tutorial, ensure that you have the Rust toolchain installed.

git clone -b writing-efficient-storage-items --single-branch
  • navigate to the project’s directory
cd sibchain

  • Run the command below to compile the node
cargo build --release

  • Run your node
./target/release/node-template --dev
  • Visit Polkadotjs to connect to the local node:

    • Go to the development tab
    • Select local node
    • Click on the switch button
  • Once the node’s UI is up, go to the extrinsic page and call the templateModule.setMemberId(memberId) extrinsic using Alice’s account with the following parameters:

    • MemberId: 1000
  • Submit the transaction

  • Now repeat the steps above for Bob’s account. (i.e., set the memberId to 1000 too).

You’ll immediately notice that there’s a problem. Multiple accounts can have the same member IDs, which is meant to be unique for each account. If you query the templateModule.idOf storage item, you’ll notice that multiple accounts have the same member ID of 1000.

  • Now try changing the memberId for Alice to 500 using the templateModule.setMemberId(memberId) extrinsic. You’ll notice that the change goes through, which isn’t the intention of setting unique member IDs (the intention is for IDs to be set once and remain unchangeable)

Solving the problem

To solve this problem, we need to introduce checks that make sure that a member cannot claim an ID that has already been taken and also to make sure that an account can’t change its ID once it has been set.

To do this:

  • Create a new storage item that keeps track of all used member IDs
  //mapping each u32 member ID to a bool. When an ID is taken, the value will be set to true
#[pallet::getter(fn taken_memberids)]
pub(super) type TakenMemberIds<T: Config> = StorageMap<_, Twox64Concat, u32, bool, ValueQuery>;
  • Write the errors that will be printed when:

    • An account has already claimed a member ID

    • A particular Member ID has been taken

      pub enum Error<T> {

      /// AccountId already exists
      AccountIdExists, //<--- Add this line

      /// Member Id taken
      MemberIdTaken, //<--- Add this line

  • Use the ensure! macro in the set_member_id dispatchable to :

    • Check if an account ID already exists in storage (in which case it will return the AccountIdExists error).

    • Check if the member ID already exists for any account (in which case it will return the MemberIdTaken error)

    • In addition to the above, mark the Member ID taken by adding the ID to the TakenMemberIds storage item and setting its value to true:

      pub fn set_member_id (origin: OriginFor<T>, member_id: u32) -> DispatchResult {

      /// Add the lines below:

      // Check if accountId already exists in storage
      ensure!(!<IdOf<T>>::contains_key(&who), Error::<T>::AccountIdExists);

      // Check if the member ID already exists for any account
      ensure!(!TakenMemberIds::<T>::contains_key(member_id), Error::<T>::MemberIdTaken);

      // Mark this member ID as taken
      TakenMemberIds::<T>::insert(member_id, true);
  • Compile and run your node:

    cargo build --release

    ./target/release/node-template --dev

You’ll now notice that the chain returns an error when you try to change the ID of an account or reassign an already-taken member ID to a different account.

Going in-depth

Substrate offers three primary types of storage items:

  1. Storage Value: Stores single data items. Ideal for system-wide configurations or flags.
  2. Storage Map: A Key-Value storage, often used for storing records indexed by account ID.
  3. Storage Double Map: Allows a two-key to value mapping. Suitable for complex nested data structures.

Storage Value

Used for global configurations. Avoid using Storage Value for user-specific data to prevent excessive storage reads, which can impact performance.

Storage Map

This is your go-to for any account-specific data. Always ensure to map to the least amount of data necessary. For example, rather than storing a complex struct, store keys to the struct and keep the elements in a separate map.

Storage Double Map

Useful for multi-dimensional data. Be wary of the computational complexity it can introduce. Always have clear documentation about the two keys to avoid misuse.

Things to take note of when declaring storage items

  • Always ensure that keys are unique. Test rigorously to ensure no key overlaps. The use of non-unique keys in storage maps can lead to data being overwritten.
  • Optimize your business logic to minimize storage operations. Frequent storage operations can lead to increased computational costs.
  • Always validate data at the runtime level using methods like ensure! to enforce business logic constraints. Not validating data before storing can lead to data integrity issues.

Best Practices for Storage Checks

Implementing storage checks usually involves using conditional statements or macros like ensure! in Rust to validate data or state conditions before proceeding with storage operations.

ensure!(!IdOf::<T>::contains_key(&who), Error::<T>::AccountIdExists);

In this example, the ensure! macro checks whether an account ID is already present in the storage map. If it is, the function returns an error, avoiding any unintended overwrites and thereby preserving data integrity.

  • Using Timeouts or Block Heights

    When dealing with actions that are time-sensitive, it's often necessary to incorporate timeouts or specific block heights as a condition.

    rustCopy code
    ensure!(<BlockNumber<T>>::get() <= expiry_block, Error::<T>::ActionExpired);

    In this case, the operation will only proceed if the current block number is less than or equal to the expiry_block, failing which, an error will be thrown.

  • Leveraging Enum States

    Sometimes, data can have multiple states. Using enums for these states and checking against them can be a prudent storage check.

    ensure!(<ContractState<T>>::get(&contract_id) == ContractState::Active, Error::<T>::ContractNotActive);

    Here, the contract will proceed only if its state is Active.

  • Composite Checks

    Sometimes, complex business logic necessitates multiple checks bundled into a single composite condition.

    rustCopy code
    ensure!(<IsUserVerified<T>>::get(&user) && <HasSufficientBalance<T>>::get(&user), Error::<T>::CannotProceed);

    Here, the action proceeds only if the user is verified and has sufficient balance.

  • Logical Checks

    Not all conditions will be straightforward true/false checks. Sometimes, logic gates (AND, OR, NOT) can make these conditions more effective.

    rustCopy code
    ensure!(!(condition_A && condition_B) || condition_C, Error::<T>::InvalidCondition);

Case Study: Ensuring Uniqueness of Member IDs

Here’s a practical example of a storage check to ensure that a member ID is unique across accounts:

rustCopy code
ensure!(!TakenMemberIds::<T>::contains_key(member_id), Error::<T>::MemberIdTaken);

Before allowing the extrinsic to set a member ID, it checks a separate TakenMemberIds storage map to ensure that the ID is unique. If it’s not, the extrinsic fails, protecting against duplicate IDs and ensuring each member has a unique identifier.


In this "Substrate in Bits" content, we have explored the challenges and solutions surrounding storage management, and best practices when dealing with storage items. Starting with a hands-on tutorial, we how multiple accounts can erroneously end up with the same member IDs. The tutorial then proceeds to introduce fixes like creating a storage item to track taken member IDs and enforcing checks to prevent ID duplication and modification.

We’ve also shed light on how to enforce data integrity and state conditions using the ensure! macro and provides several other techniques for robust storage checks. These include the use of timeouts or block heights for time-sensitive actions, leveraging Enum States for data that can exist in multiple states, and composite checks for complex business logic.

To learn more about the concepts discussed in this guide, here are some resources that we recommend:

To help us measure our progress and improve Substrate in Bits content, please fill out our living feedback form. It will only take 2 minutes of your time. Thank you!

grillchat icon