I have an array of users id's and I want to get all users with that id from the dynamoDB table
Didn't find it in the documentation
any ideas?
Add a comment
|
4 Answers
I ended up using batchGet, an operation of AWS.DynamoDB.DocumentClient
There's no support for multiple items with same key, so I have to define the key over and over again like this:
var dynamoConfig = {
sessionToken: process.env.AWS_SESSION_TOKEN,
region: process.env.AWS_REGION
};
var dynamodbDocClient = new AWS.DynamoDB.DocumentClient(dynamoConfig);
var params = {
RequestItems: {
tableName: {
Keys: [
{
id: 'user1Id'
},
{
id: 'user2Id'
},
{
id: 'user3Id'
}
]
}
}
}
dynamodbDocClient.batchGet(paramsForQueringFormerEvaluators, function(err, data) {
if (err) {
console.log('createEvaluation: get former evaluators: err: ', err);
return;
}
var users = data.Responses.tableName;
console.log('createEvaluation: get former evaluators: ', users);
});
1 Comment
Dmitry Grinko
what is paramsForQueringFormerEvaluators in your case? params?
You can use the BatchGetItem API for this.
Of course, I can't help you with any code snippet without knowing your table schema. But you can look at the documentation here.
Comments
You can probably use Scan method.
I'm currently using python boto3 package. Ref to scan in boto3. So the code would be like:
table = boto3.resourse('dynamodb').Table('users-table-name')
response = table.scan(
ScanFilter={'user-id': {
'AttributeValueList': user_ids,
'ComparisonOperator': 'IN' }
}
)
But i'm not sure what is better, using BatchGetItem or Scan. Probably it depends on two factors:
- how many records in table (
table.item_count) - how many items you need to get (
count(user_ids))
If count(user_ids) << table.item_count then you definitely need to use BatchGetItem.
If count(user_ids) / table.item_count -> 1 then you need to use Scan.
3 Comments
Andres
Avoid scan operations if you know the items id's you're looking for. They're expensive to execute because the full data set - potentially across partitions - has to be scanned.
Kloar
This can absolutely destroy you. Never use scan in a prod environment unless you have a very small table
Carlos Garces
Really bad idea, avoid use scan unless you don't known the PK. And if you don't known the PK, try to refactor your table to avoid scan!
I wrote a function that converts a list of IDs into a list of documents using batchGet:
async function getDocs(ids, docClient, tableName, key='id') {
const getParams = {
RequestItems: {
[tableName]: {
Keys: ids.map(id => ({[key]: id}))
}
}
}
return (await docClient.batchGet(getParams).promise()).Responses[tableName];
}
