Drill, mail # dev - Re: Drill terminology issue: "schema" - 2014-08-12, 04:48
Solr & Elasticsearch trainings in New York & San Francisco [more info][hide]
 Search Hadoop and all its subprojects:

Switch to Threaded View
Copy link to this message
Re: Drill terminology issue: "schema"
Regarding (4), agree with Neeraja.  Please file JIRA to this effect.

Regarding Keys' comment: workspaces are lightweight by their nature.
 Databases are generally thought of as heavyweight and managed by a dba.
This is directly in conflict with the vision of workspaces.  The purpose of
workspaces is that all users can ultimately create many workspaces for
themselves.  In many cases, they will hold views, pointers, udfs but
potentially no data.  This behavior is atypical compared to the traditional
database concept and thus is created as such.  This is part of the vision
of a loosely coupled data system.

Regarding "SCHEMA".  You're right that we have overloaded that word to mean
two entirely different things.  The first is a logical collection of
tables/views.  The latter (as in schemaless) is field names and datatypes
within a complex record.  At some level, this is the conflict with bringing
two worlds together.  Database isn't necessarily the right concept for the
former (see comments above).  However, JDBC and some BI tools use this
concept.  At the same time the latter meaning is commonly used when talking
about things like avro schemas.  I'm up for an improved approach that also
works in both domains if we can come up with one.  But let's make sure we
have good analogs before inventing something new.
On Mon, Aug 11, 2014 at 2:53 PM, Keys Botzum <[EMAIL PROTECTED]> wrote:
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB