On the order of a kilobyte, or less, per key segment is probably ideal. Beyond that, you may want to experiment to see what works for you.
Up to the order of a megabyte may work fine in newer versions (at least 1.4), but I'd start having concerns about your schema decisions after that, even if your performance was fine... since ultimately, Accumulo tables are a sorted index, and I'd wonder what kind of index needs such large keys.
I suspect a few hundred bytes is probably plenty for most applications.
Sorry for the basic question, but by key segment do you mean row key, column family column qualifier? So each should be less than a kilobyte or total? I think either way we are below that but I just want to make sure I understand the terminology. Again thanks for the reply. On Sun, Jan 26, 2014 at 2:12 AM, Christopher <[EMAIL PROTECTED]> wrote:
Yes, I mean the row, column family, column qualifier (and also column visibility). I had intended "each", not "total", because 1K is probably too constraining for some applications, but I don't think you'd see any significant difference between the two, as they are on the same order of magnitude.
Apache Lucene, Apache Solr and all other Apache Software Foundation projects and their respective logos are trademarks of the Apache Software Foundation.
Elasticsearch, Kibana, Logstash, and Beats are trademarks of Elasticsearch BV, registered in the U.S. and in other countries. This site and Sematext Group is in no way affiliated with Elasticsearch BV.
Service operated by Sematext