Jason, Thanks for the quick response. What I want to do is use a query tool that leverages some kind of Hbase secondary index to support row filtering based on columnfamily values (not rowid values). This will enable queries similar to SELECT col1 FROM table1 WHERE col1 LIKE '%test%'. Currently, this is not supported by hbase, AFAIK.
Now, it is possible to create Hbase secondary indexes using the classes in org.apache.hadoop.hbase.client.tableindexed (see SingleColumnValueFilter). This approach works and I can retrieve filter results using a basic java client. The problem is that there is nothing layered on top of the client. I need something like a JavaEE app with a web front end where a user could enter a query text string and leverage the above java client code to retrieve filtered rows from bhase. Now, I have implemented a solution using Solr/Lucene as the indexing and query server[1]. See below for details. The downside is my implementation involves two queries: one to the Solr index store and a second one to hbase. I would prefer just one query directly to hbase. I hope this makes sense. Please let me know if your implementation seems to fit my requirements. I'll continue to read about what you have done. Paul [1] What I am looking for is close to this: https://issues.apache.org/jira/browse/SOLR-1301?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel This enables the creation of Hbase secondary indexes via Solr/Lucene. The Solr indexes are stored in Solr's native storage format. I have used this module to index Hbase rows with the row id mapped to the Solr Document id. Similarly, Hbase column families are mapped to Solr fields. This enables someone to lookup an Hbase column using a term query such as TransactionInfo.transaction_nbr:9043. I have also created an implementation of a org.apache.solr.response.QueryResponseWriter. This class receives the results of the Solr query, looks up the Solr Documents, retrieves the all ids which map to the Hbase row ids, sorts them, and then executes a range scan on the Hbase table using startRow and stopRow. This works reasonably well but it involves two queries: one to Solr and one to Hbase. -----Original Message----- From: Jason Rutherglen [mailto:[email protected]] Sent: Friday, May 13, 2011 12:00 PM To: [email protected] Subject: Re: Web-based Hbase Query Tool? Paul, I have a patch to implement Lucene in HBase - HBASE-3529. It's a work in progress limited primarily by time. For a web UI I'll probably eventually integrate Solr at some point, which would provide a basic UI, though it'd need to be modified to account for column families. Is this what you're looking for? Jason On Fri, May 13, 2011 at 8:52 AM, Sterk, Paul (Contractor) <[email protected]> wrote: > Hi, > > > > I am looking for a web based query tool that is able to query Hbase > tables and columns. Can someone point me to any? I want to be able to > create secondary indexes on Hbase tables then issue queries against > similar to Solr/Lucene. For example: 'find all transactions with > transaction number 9040' with a query such as > 'TransactionMaster.transaction_nbr:9040'. > > > > I have found this tool so far: > http://sourceforge.net/projects/hbaseexplorer/ > > > > Please let me know if there are any others. > > > > Thanks in advance, > > Paul > > This message, including any attachments, is the property of Sears Holdings > Corporation and/or one of its subsidiaries. It is confidential and may > contain proprietary or legally privileged information. If you are not the > intended recipient, please delete it without reading the contents. Thank you. > This message, including any attachments, is the property of Sears Holdings Corporation and/or one of its subsidiaries. It is confidential and may contain proprietary or legally privileged information. If you are not the intended recipient, please delete it without reading the contents. Thank you.
