Dynamodb: Querying tables with secondary index

Question

I am using the gem aws-sdk-ruby to query a table which looks something like this:

hk (Hashkey)  |  guid(Rangekey)  |  Timestamp (Secondary Range index)  |  other attributes
aaaa          |  50              |  2013-02-04T12:33:00Z               |
aaaa          |  244             |  2013-04-22T04:54:00Z               |
aaaa          |  342             |  2013-05-18T06:52:00Z               |
bbbb          |  243             |  2013-06-21T13:17:00Z               |

What I am trying to do is get all 'aaaa' rows that were created after a certain date. Ex:

AWS.config(access_key_id: 'xxx', secret_access_key: 'xxx', :dynamo_db => { :api_version => '2012-08-10' })
client = AWS::DynamoDB::Client.new
client.query(
{
  table_name: 'table',
  select: 'ALL_ATTRIBUTES',
  key_conditions: {
    'hk' => {
      comparison_operator: 'EQ',
      attribute_value_list: [
        {'s' => 'aaaa'}
      ]
    },
    'timestamp' => {
      comparison_operator: 'GE',
      attribute_value_list: [
        {'s' => Time.now.utc.iso8601}
      ]
    }
  }
})

When i run the code above i get this:

Query condition missed key schema element guid (AWS::DynamoDB::Errors::ValidationException)

Running the query with hashKey and RangeKey works but when i replace the rangeKey with a secondary range index it fails telling me that the rangeKey is required.

If i then add the range key (Which makes no sense) i get the following error instead:

Conditions can be of length 1 or 2 only (AWS::DynamoDB::Errors::ValidationException)

Anyone know what might be happening?

Answer 1

You are not querying the secondary index, you are querying the primary index(hash and range key). To use a secondary index with DynamoDB you must use V2 of the API and specify the index in the query operation

client = AWS::DynamoDB::Client.new(api_version: '2012-08-10') 

client.query( {   :table_name: 'table',   :index_name: "timestamp-index", :select: 'ALL_PROJECTED_ATTRIBUTES',   :key_conditions: {
    'hk' => {
      :comparison_operator: 'EQ',
     :attribute_value_list: [
        {'s' => 'aaaa'}
      ]
    },
    'timestamp' => {
     :comparison_operator: 'GE',
      :attribute_value_list: [
        {'s' => Time.now.utc.iso8601}
      ]
    }   } })

Answer 2

You can use "scan" instead "query" http://docs.aws.amazon.com/AWSJavaScriptSDK/latest/AWS/DynamoDB/DocumentClient.html#scan-property

I can emphatize that scan in slower than query, scan is not recomended for huge amount data retrieve.

I did this with the next schema

DB Schema Where Primary partition key id (String)

  //JavaScript EXAMPLE-1  
var params = {
        TableName : 'users',
        FilterExpression : 'isActive = :isActive',
        ExpressionAttributeValues : {':isActive' : true}
    };


        dynamoDBClient.scan(params, function(err, data){
            if(err){
                return console.error(err);
            }
            console.log(data.Items);

        });

I hope help you.

Regards.