Hi Lars, That's an interesting observation, I haven't thought before about scans in HAcid. Your suggestion for a solution is really close to what I would do: implement HAcidScan as a HBase Scan that filters according to the cache column.
Thanks, I will add this feature to the to do list. -- Andre Medeiros ________________________________________ From: lars hofhansl [[email protected]] Sent: Tuesday, October 02, 2012 03:53 To: [email protected] Cc: [email protected] Subject: Re: HAcid: multi-row transactions in HBase Interesting stuff Andre. I noticed you do not deal with scans. I suppose this is because it is hard to do scans touching many rows efficiently with client only code. This can probably be handled with an HBase Filter on the Scan, which filters as much as it can based on the cache you maintain in the per-table column and then post filter the remaining rows at the client (by inspecting the global TimestampTable). -- Lars ----- Original Message ----- From: de Souza Medeiros Andre <[email protected]> To: "[email protected]" <[email protected]> Cc: Sent: Monday, October 1, 2012 4:12 PM Subject: HAcid: multi-row transactions in HBase Hello all at this mailing list, I'm glad to finally announce an HBase addon that I have been working on. HAcid is a client library that enables HBase applications to execute ACID transactions with Snapshot Isolation. It's purpose is similar to Google Percolator for Bigtable. The library is open-source under Apache 2 license. http://staltz.blogspot.fi/2012/10/hacid-multi-row-transactions-in-hbase.html It might still be rough around the edges, so please let me know about bugs and issues (e.g., in the bitbucket repo). And let me know if HAcid was somehow useful to you. Thanks, Andre Medeiros http://about.me/andrestaltz
