ApsaraDB for HBase:Use aliases

Background information

1. In the business logic, Table A is configured as the destination table. If you need to change the destination table to Table B, you must modify the configuration.
2. Table sharding

   In most business scenarios, only data of the previous week is frequently accessed. You can create a new table each week to store the latest data and efficiently query hot data. However, you must manually create and delete tables. This increases the complexity of business.

   This topic describes how to use aliases to resolve the preceding issues.

Scenarios

Manage time series data

If your business data is time-sensitive, multiple indexes can be created based on time. This reduces the size of a single index and improves the query performance. During this process, you do not need to manually create or delete indexes.

Rebuild an index

You can rebuild indexes. The rebuild operation does not affect existing index-based data queries. After the indexes are rebuilt, the new indexes are used to query data. During this process, you do not need to modify any existing code.

Note: The following section describes how to run curl commands to use aliases.

Instructions

Basic feature: Create an alias that points to an existing collection

curl "http://solrhost:8983/solr/admin/collections?action=CREATEALIAS&name=your_alias_name&collections=your_collection_name_A"

The preceding curl command creates an alias named your_alias_name. The alias points to the collection your_collection_name_A. You can set your_alias_name as the destination in the business logic. Then, the kernel forwards query requests to the collection. If you want to change the name of the collection to your_collection_name_B, run the curl command to change the alias.

Change the alias

curl "http://solrhost:8983/solr/admin/collections?
action=ALIASPROP&name=your_alias_name&collections=your_collection_name_B"

This way, you can query data in the new collection without the need to modify your business code.

Advanced feature: Automatically create collections

The Search service can automatically create collections based on the time field. This simplifies your business logic. The following example shows how to configure the system to create a collection every week and automatically delete the collections that have expired.

curl "http://solrhost:8983/solr/admin/collections?action=CREATEALIAS&name=test_router_alias&router.start=NOW-30DAYS/DAY&router.autoDeleteAge=/DAY-90DAYS&router.field=your_timestamp_l&router.name=time&router.interval=%2B7DAY&router.maxFutureMs=8640000000&create-collection.collection.configName=_indexer_default&create-collection.numShards=2"

The preceding curl command creates collections every seven days based on the value of the your_timestamp_l field. The beginning of the time range for the first collection is 30 days before. The difference between the value of the your_timestamp_l field and the current time is within 100 days. The collections that were created 90 days before are automatically deleted.

Precautions
1. The business data must contain a time field. The data type of the time field can be DATE or LONG.
2. By default, all collections are queried. You can specify a collection to query. To do this, you must run the curl command with the specified URL or call the specific API operation to obtain all collections. Then, you can obtain the time field from the collections and determine the collection to access. For more information, see the following sample code.

Delete an alias

To delete a common alias, run the following command:

curl "http://solrhost:8983/solr/admin/collections?action=DELETEALIAS&name=your_alias_name"

If an alias is used to automatically create collections, you must delete the collections after you delete the alias. To delete the alias and collections, perform the following steps:

1. Query the collections that are created by using the alias.

   curl "http://solrhost:8983/solr/admin/collections?action=LIST"

   In this example, all the collections whose names start with test_router_alias are created by using the alias.
2. Delete the alias.

   curl "http://solrhost:8983/solr/admin/collections?action=DELETEALIAS&name=test_router_alias"

3. Delete the collections.

   curl "http://solrhost:8983/solr/admin/collections?action=DELETE&name=collection_name"

Reference

https://lucene.apache.org/solr/guide/7_3/collections-api.html#createalias

https://lucene.apache.org/solr/guide/7_3/collections-api.html#list

Find a collection based on a time field of the LONG type

import org.apache.solr.client.solrj.SolrQuery;
import org.apache.solr.client.solrj.impl.CloudSolrClient;
import org.apache.solr.client.solrj.impl.ClusterStateProvider;
import org.apache.solr.client.solrj.response.QueryResponse;
import org.apache.solr.common.SolrDocument;
import org.apache.solr.common.util.StrUtils;

import java.time.Instant;
import java.time.ZoneOffset;
import java.time.format.DateTimeFormatter;
import java.time.format.DateTimeFormatterBuilder;
import java.time.temporal.ChronoField;
import java.util.AbstractMap;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Locale;
import java.util.Map;
import java.util.Optional;

public class SolrDemo {
  private static final DateTimeFormatter DATE_TIME_FORMATTER = new DateTimeFormatterBuilder()
      .append(DateTimeFormatter.ISO_LOCAL_DATE).appendPattern("[_HH[_mm[_ss]]]")
      .parseDefaulting(ChronoField.HOUR_OF_DAY, 0)
      .parseDefaulting(ChronoField.MINUTE_OF_HOUR, 0)
      .parseDefaulting(ChronoField.SECOND_OF_MINUTE, 0)
      .toFormatter(Locale.ROOT).withZone(ZoneOffset.UTC);

  private static final String zkHost = "localhost:2181/solr";
  private CloudSolrClient cloudSolrClient;
  private ClusterStateProvider clusterStateProvider;

  public SolrDemo() {
    cloudSolrClient = new CloudSolrClient.Builder(
        Collections.singletonList(zkHost), Optional.empty()).build();
    cloudSolrClient.connect();
    clusterStateProvider = cloudSolrClient.getClusterStateProvider();
  }

  public void close() throws Exception {
    if (null != cloudSolrClient) {
      cloudSolrClient.close();
    }
  }

  private List<String> findCollection(String aliasName, long start, long end) {
    List<String> collections = new ArrayList<>();
    if (start > end) {
      return collections;
    }
    // Query collections based on the time range specified by the start and end parameters.
    if (clusterStateProvider.getState(aliasName) == null) {
      // Query all the collections that are created by using the alias specified by the aliasName parameter.
      // test_router_alias_2020-03-04, test_router_alias_2020-02-26, test_router_alias_2020-02-19, test_router_alias_2020-02-12, test_router_alias_2020-02-05
      List<String> aliasedCollections = clusterStateProvider.resolveAlias(aliasName);

      // Extract date and time from the name of each collection.
      // 2020-03-04T00:00:00Z=test_router_alias_2020-03-04,
      // 2020-02-26T00:00:00Z=test_router_alias_2020-02-26,
      // 2020-02-19T00:00:00Z=test_router_alias_2020-02-19,
      // 2020-02-12T00:00:00Z=test_router_alias_2020-02-12,
      // 2020-02-05T00:00:00Z=test_router_alias_2020-02-05
      List<Map.Entry<Instant, String>> collectionsInstant = new ArrayList<>(aliasedCollections.size());
      for (String collectionName : aliasedCollections) {
        String dateTimePart = collectionName.substring(aliasName.length() + 1);
        Instant instant = DATE_TIME_FORMATTER.parse(dateTimePart, Instant::from);
        collectionsInstant.add(new AbstractMap.SimpleImmutableEntry<>(instant, collectionName));
      }

      // Find the required collection based on the query time.
      Instant startI = Instant.ofEpochMilli(start);
      Instant endI = Instant.ofEpochMilli(end);
      for (Map.Entry<Instant, String> entry : collectionsInstant) {
        Instant colStartTime = entry.getKey();
        if (! endI.isBefore(colStartTime)) {
          collections.add(entry.getValue());
          System.out.println("find collection: " + entry.getValue());
          if (! startI.isBefore(colStartTime)) {
            break;
          }
        }
      }
    } else {
      collections.add(aliasName);
    }
    System.out.println("query " + collections);
    return collections;
  }

  public void run() throws Exception {
    try {
      // [2020-03-07 2020-03-10]
      long start = 1583538686312L;
      long end = 1583797886000L;
      String aliasName = "test_router_alias";
      String collections = StrUtils.join(findCollection(aliasName, start, end), ',');
      QueryResponse res = cloudSolrClient.query(collections, new SolrQuery("*:*"));
      for (SolrDocument sd : res.getResults()) {
        System.out.println(sd.get("id") + " " + sd.get("gmtCreate_l"));
      }
    } finally {
      cloudSolrClient.close();
    }
  }

  public static void main(String[] args) throws Exception {
    SolrDemo solrDemo = new SolrDemo();
    solrDemo.run();
    solrDemo.close();
  }
}