In previous posts we have already spoken of Apache Ignite and its capabilities as a distributed SQL/Lucene database.

This feature allows us to analyze information stored in our cluster in a simple way, using SQL and Lucene expressions. However, the standard functionality provided by Apache Ignite may not be sufficient in certain circumstances:

  • When you need immediate availability of the cluster when using Lucene or geospatial indexing with native persistence.

    • The standard Lucene and geospatial indexes provided by Apache Ignite are not available after a reboot, as they do not support native disk persistence. This feature forces a manual re-indexing, with the consequent delay in the availability of the database cluster.

  • When you need to perform complex searches.

    • The standard Lucene indexing module provided by Apache Ignite only supports textual searches.

    • Performing complex searches using only standard SQL can be an almost impossible task, with a huge impact on performance.

      For example, think about how to search for geo-positioned tweets written from the U.S. and Europe, between two dates containing the #happy or #feliz hashtags, in a database with billions of entries, with a single query and that the answer is available in milliseconds.

  • When you need to change your database schema.

    • These changes can only be made manually and the creation of the indexes is atomic, i.e., the indexes must be created one by one and for each index created a complete scan of the data is performed. Think of a database with billions of records and the performance penalty that goes with it.

To remove these limitations, at Hawkore we have developed a set of modules that extend the search, indexing, persistence, and schema mutation capabilities of Apache Ignite’s distributed SQL/Lucene database.

advanced-indexing

  • Instant availability of a reliable database cluster after a reboot. Our Apache Lucene and Geospatial indexes support optimized disk persistence, without losing the high performance of the distributed in-memory computing layer.

  • Combines the SQL capabilities of Apache Ignite with extended Apache Lucene. Yes, you read it well! You can do a SQL query using Lucene filtering expressions that can be as complex as you want, with a response time of milliseconds.

  • You can modify the Apache Ignite database schemas (QueryEntity) at runtime without needing a reboot and without loss of information:

    • Add new fields (indexed or not) to your entities at runtime.
    • Optimized index reconstruction process which automatically detects if re-indexing is necessary, and can do it simultaneously for all new indexes with only one background data scan.

Main features of the extended Lucene index

  • Supports disk persistence, without losing the performance of the in-memory computing layer, allowing your data cluster to be immediately available.
  • Secure implementation for Apache Ignite EncryptionSpi since the Lucene index does not store the cache entry value.
  • Supports updating and rebuilding of indexes at runtime.
  • Supports TextQuery with Lucene’s advanced search expressions.
  • Supports the syntax of the classic Lucene query analyzer in the Lucene search expression.
  • Full-text search (language recognition analysis, wildcard, diffuse, regular expression).
  • Boolean search (and, or, not).
  • Sorting by relevance, column value, and distance.
  • Geospatial indexing (points, lines, polygons, and their collections).
  • Geospatial transformations (bounding box, buffer, centroid, convex hull, union, difference, and intersection).
  • Geospatial operations (intersection, contains, is in).
  • Bitemporal search (valid time and transaction time).
  • Paging across cluster, even with ordered searches.
  • Integration with Spring Data 2.0 SpEL.

You want to know how it works?

Sign up at www.hawkore.com, access all the documentation with practical examples and try it out for Free!