Skip to main content

Data Storage / Collections

All data stored on the NEAR blockchain is done in key / value pairs. There are several collection methods in the SDKs we've created that will help you store your data on chain.

For information on storage costs, please see [ storage staking ].


Rust Collection Typesโ€‹

near-sdk-rs module documentation

TypeIterableClear All ValuesPreserves Insertion OrderRange Selection
Vectorโœ…โœ…โœ…โœ…
LookupSet
UnorderedSetโœ…โœ…โœ…
LookupMap
UnorderedMapโœ…โœ…โœ…
TreeMapโœ…โœ…

Big-O Notationโ€‹

The Big-O notation values in the chart below describe the time complexity of the various collection methods found in near-sdk-rs. These method complexities correlate with gas consumption on NEAR, helping you decide which collection to utilize in your project. There are three types found in our collection methods:

TypeAccessInsertDeleteSearchTraverseClear
VectorO(1)O(1)*O(1)**O(n)O(n)O(n)
LookupSetO(1)O(1)O(1)O(1)N/AN/A
UnorderedSetO(1)O(1)O(1)O(1)O(n)O(n)
LookupMapO(1)O(1)O(1)O(1)N/AN/A
UnorderedMapO(1)O(1)O(1)O(1)O(n)O(n)
TreeMapO(1)O(log n)O(log n)O(log n)O(n)O(n)

* - to insert at the end of the vector using push_back (or push_front for deque)

** - to delete from the end of the vector using pop (or pop_front for deque), or delete using swap_remove which swaps the element with the last element of the vector and then removes it.


Gas Consumption Examplesโ€‹

The examples below show differences in gas burnt storing and retrieving key/value pairs using the above methods. Please note that the gas cost of spinning up the runtime environment on chain has been deducted to show just data read/writes.

You can reproduce this and test out your own data set by visiting collection-examples-rs.

Rust Set Data Gas Chart

Rust Get Data Gas Chart


Vectorโ€‹

Implements a vector / persistent array.

  • can iterate using index
  • Uses the following map: index -> element.

[ SDK source ]

[ Implementation ]


LookupSetโ€‹

Implements a persistent set without iterators.

  • can not iterate over keys
  • more efficient in reads / writes

[ SDK source ]

[ Implementation ]


UnorderedSetโ€‹

Implements a persistent set with iterators for keys, values, and entries.

[ SDK source ]

[ Implementation Docs ]


LookupMapโ€‹

Implements a persistent map.

  • can not iterate over keys
  • does not preserve order when removing and adding values
  • efficient in number of reads and writes
  • To add data:
pub fn add_lookup_map(&mut self, key: String, value: String) {
self.lookup_map.insert(&key, &value);
}
  • To get data:
pub fn get_lookup_map(&self, key: String) -> String {
match self.lookup_map.get(&key) {
Some(value) => {
let log_message = format!("Value from LookupMap is {:?}", value.clone());
env::log(log_message.as_bytes());
value
},
None => "not found".to_string()
}
}

[ SDK source ]

[ Implementation ]


UnorderedMapโ€‹

Implements an unordered map.

  • iterable
  • does not preserve order when removing and adding values
  • is able to clear all values
  • To add data:
pub fn add_unordered_map(&mut self, key: String, value: String) {
self.unordered_map.insert(&key, &value);
}
  • To get data:
pub fn get_unordered_map(&self, key: String) -> String {
match self.unordered_map.get(&key) {
Some(value) => {
let log_message = format!("Value from UnorderedMap is {:?}", value.clone());
env::log(log_message.as_bytes());
value
},
// None => "Didn't find that key.".to_string()
None => "not found".to_string()
}
}

[ SDK source ]

[ Implementation ]


TreeMapโ€‹

Implements a Tree Map based on AVL-tree.

  • iterable
  • preserves order
  • able to clear all values
  • self balancing
  • To add data:
pub fn add_tree_map(&mut self, key: String, value: String) {
self.tree_map.insert(&key, &value);
}
  • To get data:
pub fn get_tree_map(&self, key: String) -> String {
match self.tree_map.get(&key) {
Some(value) => {
let log_message = format!("Value from TreeMap is {:?}", value.clone());
env::log(log_message.as_bytes());
// Since we found it, return it (note implicit return)
value
},
// did not find the entry
// note: curly brackets after arrow are optional in simple cases, like other languages
None => "not found".to_string()
}
}

[ SDK source ]

[ Implementation ]


Storage Constraints on NEARโ€‹

For storing data on-chain itโ€™s important to keep in mind the following:

  • Can add up in storage staking costs
  • There is a 4mb limit on how much you can upload at once

Letโ€™s say for example, someone wants to put an NFT purely on-chain (rather than IPFS or some other decentralized storage solution) youโ€™ll have almost an unlimited amount of storage but will have to pay 1 $NEAR per 100kb of storage used (see Storage Staking).

Users will be limited to 4MB per contract call upload due to MAX_GAS constraints. The maximum amount of gas one can attach to a given functionCall is 300TGas.

Was this page helpful?