ElasticSearch-Hadoop: Indexingproductviews count andcustomer topsearch queryfrom Hadoop to ElasticSe

原文地址:http://www.javacodegeeks.com/2014/05/elasticsearch-hadoop-indexing-product-views-count-and-customer-top-search-query-from-hadoop-to-elasticsearch.html

This post covers to use ElasticSearch-Hadoop to read data from Hadoop system and index that in ElasticSearch. The functionality it covers is to index product views count and top search query per customer in last n number of days. The analyzed data can further be used on website to display customer recently viewed, product views count and top search query string.

In continuation to the previous posts on

•  Customer product search clicks analytics using big data,
•  Flume: Gathering customer product search clicks data using Apache Flume,
•  Hive: Query customer top search query and product views count using Apache Hive.

We already have customer search clicks data gathered using Flume and stored in Hadoop HDFS and ElasticSearch, and how to analyze same data using Hive and generate statistical data. Here we will further see how to use the analyzed data to enhance customer experience on website and make it relevant for the end customers.

Recently Viewed Items

We already have covered in first part, how we can use flume ElasticSearch sink to index the recently viewed items directory to ElasticSearch instance and the data can be used to display real time clicked items for the customer.

ElasticSearch-Hadoop

Elasticsearch for Apache Hadoop allows Hadoop jobs to interact with ElasticSearch with small library and easy setup.

Elasticsearch-hadoop-hive, allows to access ElasticSearch using Hive. As shared in previous post, we have product views count and also customer top search query data extracted in Hive tables. We will read and index the same data to ElasticSearch so that it can be used for display purpose on website.

Product views count functionality

Take a scenario to display each product total views by customer in the last n number of days. For better user experience, you can use the same functionality to display to end customer how other customer perceive the same product.

Hive Data for product views

Select sample data from hive table:

1# search.search_productviews : id, productid, viewcount

261, 61, 15

348, 48, 8

416, 16, 40

585, 85, 7

Product Views Count Indexing

Create Hive external table “search_productviews_to_es” to index data to ElasticSearch instance.

1Use search;

2DROPTABLE IF EXISTS search_productviews_to_es;

3CREATEEXTERNAL TABLEsearch_productviews_to_es (id STRING, productid BIGINT, viewcount INT) STORED BY'org.elasticsearch.hadoop.hive.EsStorageHandler' TBLPROPERTIES('es.resource'= 'productviews/productview','es.nodes'= 'localhost','es.port'= '9210','es.input.json'= 'false','es.write.operation'= 'index','es.mapping.id'= 'id','es.index.auto.create'= 'yes');

4INSERTOVERWRITE TABLEsearch_productviews_to_es SELECTqcust.id, qcust.productid, qcust.viewcount FROMsearch_productviews qcust;

•  External table search_productviews_to_es is created points to ES instance
•  ElasticSearch instance configration used is localhost:9210
•  Index “productviews” and document type “productview” will be used to index data
•  Index and mappins will automatically created if it does not exist
•  Insert overwrite will override the data if it already exists based on id field.
•  Data is inserting by selecting data from another hive table “search_productviews” storing analytic/statistical data.

Execute the hive script in java to index product views data, HiveSearchClicksServiceImpl.java

1Collection<HiveScript> scripts = newArrayList<>();

2            HiveScript script = newHiveScript(newClassPathResource("hive/load-search_productviews_to_es.q"));

3            scripts.add(script);

4            hiveRunner.setScripts(scripts);

5            hiveRunner.call();

productviews index sample data

The sample data in ElasticSearch index is stored as below:

1{id=48, productid=48, viewcount=10}

2{id=49, productid=49, viewcount=20}

3{id=5, productid=5, viewcount=18}

4{id=6, productid=6, viewcount=9}

Customer top search query string functionality

Take a scenario, where you may want to display top search query string by a single customer or all the customers on the website. You can use the same to display top search query cloud on the website.

Hive Data for customer top search queries

Select sample data from hive table:

1# search.search_customerquery : id, querystring, count, customerid

261_queryString59, queryString59, 5, 61

3298_queryString48, queryString48, 3, 298

4440_queryString16, queryString16, 1, 440

547_queryString85, queryString85, 1, 47

Customer Top search queries Indexing

Create Hive external table “search_customerquery_to_es” to index data to ElasticSearch instance.

1Use search;

2DROPTABLE IF EXISTS search_customerquery_to_es;

3CREATEEXTERNAL TABLEsearch_customerquery_to_es (id String, customerid BIGINT, querystring String, querycount INT) STORED BY'org.elasticsearch.hadoop.hive.EsStorageHandler' TBLPROPERTIES('es.resource'= 'topqueries/custquery','es.nodes'= 'localhost','es.port'= '9210','es.input.json'= 'false','es.write.operation'= 'index','es.mapping.id'= 'id','es.index.auto.create'= 'yes');

4INSERTOVERWRITE TABLEsearch_customerquery_to_es SELECTqcust.id, qcust.customerid, qcust.queryString, qcust.querycount FROMsearch_customerquery qcust;

•  External table search_customerquery_to_es is created points to ES instance
•  ElasticSearch instance configration used is localhost:9210
•  Index “topqueries” and document type “custquery” will be used to index data
•  Index and mappins will automatically created if it does not exist
•  Insert overwrite will override the data if it already exists based on id field.
•  Data is inserting by selecting data from another hive table “search_customerquery” storing analytic/statistical data.

Execute the hive script in java to index data HiveSearchClicksServiceImpl.java

1Collection<HiveScript> scripts = newArrayList<>();

2            HiveScript script = newHiveScript(newClassPathResource("hive/load-search_customerquery_to_es.q"));

3            scripts.add(script);

4            hiveRunner.setScripts(scripts);

5            hiveRunner.call();

topqueries index sample data

The topqueries index data on ElasticSearch instance is as shown below:

1{id=474_queryString95, querystring=queryString95, querycount=10, customerid=474}

2{id=482_queryString43, querystring=queryString43, querycount=5, customerid=482}

3{id=482_queryString64, querystring=queryString64, querycount=7, customerid=482}

4{id=483_queryString6, querystring=queryString6, querycount=2, customerid=483}

5{id=487_queryString86, querystring=queryString86, querycount=111, customerid=487}

6{id=494_queryString67, querystring=queryString67, querycount=1, customerid=494}

The functionality described above is only sample functionality and ofcourse need to be extended to map to specific business scenario. This may cover business scenario of displaying search query cloud to customers on website or for further Business Intelligence analytics.

Spring Data

Spring ElasticSearch for testing purpose has also been included to create ESRepository to count total records and delete All.
Check the service for details, ElasticSearchRepoServiceImpl.java

Total product views:

01@Document(indexName = "productviews", type = "productview", indexStoreType = "fs", shards = 1, replicas = 0, refreshInterval = "-1")

02publicclass ProductView {

03    @Id

04    privateString id;

05    @Version

06    privateLong version;

07    privateLong productId;

08    privateint viewCount;

09    ...

10    ...

11    }

12 

13publicinterface ProductViewElasticsearchRepository extendsElasticsearchCrudRepository<ProductView, String> { }

14 

15longcount = productViewElasticsearchRepository.count();

Customer top search queries:

01@Document(indexName = "topqueries", type = "custquery", indexStoreType = "fs", shards = 1, replicas = 0, refreshInterval = "-1")

02publicclass CustomerTopQuery {

03    @Id

04    privateString id;

05    @Version

06    privateLong version;

07    privateLong customerId;

08    privateString queryString;

09    privateint count;

10    ...

11    ...

12    }

13 

14publicinterface TopQueryElasticsearchRepository extendsElasticsearchCrudRepository<CustomerTopQuery, String> { }

15 

16longcount = topQueryElasticsearchRepository.count();

In later posts we will cover to analyze the data further using scheduled jobs,

• Using Oozie to schedule coordinated jobs for hive partition and bundle job to index data to ElasticSearch.
• Using Pig to count total number of unique customers etc.