Table of Contents

 
 
 

Summary

Lucene.NET is a .NET implementation of the Lucene full-text search engine. This article describes how you can use Lucene.NET to index text data stored in Windows Azure SQL Database, and then perform searches against that data.
NOTE: This does not provide an integrated full-text search experience like the full-text search in SQL Server. Lucene.NET is an external process that is queried separately from SQL Database.
NOTE: This article relies on the Azure Library for Lucene.NET (https://azuredirectory.codeplex.com/ ) to store the Lucene.NET catalog in a Windows Azure storage blob.
Prerequisites

To use the Azure Library for Lucene.NET and Lucene.NET from a Visual Studio project, you must add a reference to both the AzureDirectory project or assembly, and the Lucerne.NET project or assembly.  You must also add the following using statements to your project: using Lucene.Net; using Lucene.Net.Store; using Lucene.Net.Index; using Lucene.Net.Search; using Lucene.Net.Documents; using Lucene.Net.Util; using Lucene.Net.Analysis; using Lucene.Net.Analysis.Standard; using Lucene.Net.Search; using Lucene.Net.QueryParsers; using Lucene.Net.Store.Azure;

Lucene.NET

Lucene.NET is a .NET implementation of Lucene (http://lucene.apache.org/ ) and provides full-text indexing and search of documents.  Documents are composed of multiple fields and do not have a predefined schema. When performing a query against the index, you can search across multiple fields within a document. Lucene.NET doesn't directly integrate with SQL Database; instead you must perform a query against a database and construct a Document from the results, which is then cataloged by Lucene.Net.  For more information on Lucene.NET, see http://lucenenet.apache.org/ .

The Azure Library for Lucene.NET

This library allows you to expose blob storage as a Lucene.NET.Store.Directory object, which Lucene.NET uses as persistent storage for its catalog. More information on the Azure Library for Lucene.NET, as well as the latest version, can be found on the project homepage at https://azuredirectory.codeplex.com/ .
The current version of the Azure Library (as of 22 May 2013) may require modification before using it in your solution. Specifically:

  • It may launch a the Visual Studio project conversion wizard when launched.
  • The reference to Microsoft.WindowsAzure.Storage may need to be deleted and recreated to point to the most recent version of the assembly.
  • There are several Debug.WriteLine statements that should be converted to Trace.write or another member of the Trace class as documented at http://msdn.microsoft.com/en-us/library/ff966484.aspx . If you are not interested in diagnostic output, you can simply remove the Debug.WriteLine statements.

Using the Library

The following code creates an AzureDirectory object and uses it as a parameter when creating the IndexWriter:

AzureDirectory azureDirectory = new AzureDirectory(
        CloudStorageAccount.FromConfigurationSetting(
        "Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString"), "TestCatalog");
  
IndexWriter indexWriter = new IndexWriter(azureDirectory, new StandardAnalyzer(), true);

Using Lucene.NET to index SQL Database

As mentioned previously, Lucene.NET is not integrated directly with SQL Database and is based on indexing 'documents' that contain multiple fields.  In order to index data from SQL Database, you must query the database and create a new Document object for each row. Individual columns can then be added to the Document. The following code illustrates querying a SQL Database that contains information on individual bloggers, and then adding the ID and Bio column information to the Lucene index using an IndexWriter and Document:

// Create the AzureDirectory against blob storage and create a catalog named 'Catalog'
AzureDirectory azureDirectory= new AzureDirectory(CloudStorageAccount.FromConfigurationSetting("Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString"), "Catalog");
  
IndexWriter indexWriter = new IndexWriter(azureDirectory, new StandardAnalyzer(), true);
indexWriter.SetRAMBufferSizeMB(10.0);
indexWriter.SetUseCompoundFile(false);
indexWriter.SetMaxMergeDocs(10000);
indexWriter.SetMergeFactor(100);
  
// Create a DataSet and fill it from SQL Database
DataSet ds = new DataSet();
using(SqlConnection sqlCon = new SqlConnection(sqlConnString))
{  
    sqlCon.Open();  
    SqlCommand sqlCmd = new SqlCommand();  
    sqlCmd.Connection = sqlCon;  
    sqlCmd.CommandType = CommandType.Text;  
    // Only get the minimum fields we need; Bio to index, Id so search results
    // can look up the record in SQL Database
    sqlCmd.CommandText = "select Id, Bio from bloggers";  
    SqlDataAdapter sqlAdap = new SqlDataAdapter(sqlCmd);  
    sqlAdap.Fill(ds);  
}  
if (ds.Tables[0] != null)  
{  
    DataTable dt = ds.Tables[0];  
    if (dt.Rows.Count > 0)  
    {  
        foreach (DataRow dr in dt.Rows)  
        {  
            // Create the Document object
            Document doc = new Document();  
            foreach (DataColumn dc in dt.Columns)
            {
                // Populate the document with the column name and value from our query
                doc.Add(new Field(
                       dc.ColumnName, 
                       dr[dc.ColumnName].ToString(), 
                       Field.Store.YES, 
                       Field.Index.TOKENIZED));
            }
            // Write the Document to the catalog
            indexWriter.AddDocument(doc);  
        }  
    }  
}  
// Close the writer
indexWriter.Close();

Note: The above sample returns all rows and adds them to the catalog. In a production application you will most likely only want to add new or updated rows.

Searching the Lucene.NET catalog

After you have added documents to the catalog, you can perform a search against them using the IndexSearcher. The following example illustrates how to create perform a search against the catalog for a term contained in the 'Bio' field and return the Id of that result:

// Create the AzureDirectory for blob storage
AzureDirectory azureDirectory = new AzureDirectory(CloudStorageAccount.FromConfigurationSetting("Microsoft.WindowsAzure.Plugins.Diagnostics.ConnectionString"), "Catalog");
// Create the IndexSearcher
IndexSearcher indexSearcher = new IndexSearcher(azureDirectory);
// Create the QueryParser, setting the default search field to 'Bio'
QueryParser parser = new QueryParser("Bio", new StandardAnalyzer());
// Create a query from the Parser
Query query = parser.Parse(searchString);
// Retrieve matching hits 
Hits hits = indexSearcher.Search(query);
// Loop through the matching hits, retrieving the document
for (int i = 0; i < hits.Length(); i++)
{
    //Retrieve the string value of the 'Id' field from the
    //hits.Doc(i) document.
    TextBox_Results.Text += "Id: " + hits.Doc(i).GetField("Id").StringValue()+ "\n";
}

Based on the Id, you can perform a query against SQL Database to return additional fields from the matching record.

References