elastic map reduce - Can BigQuery's browser interface be white-labeled? -
like people, we're pretty impressed bigquery. we're willing put being based on proprietary "dremel" in exchange not having configure ton of servers in our lan, on ec2, or anywhere else.
the rest api excellent, , we're incorporating our apps, still find ourselves using bq browser interface well. we'd incorporate 'generic sql window' our app, without divulging backend bq or data stored in google @ all, matter. does google provide way use bq browser tool in white-label manner?
note also, extending access existing browser tool problematic. relies on user-accounts existing in one's own domain - can't done, in our case, customer's email address. rest interface solves service-level accounts, doesn't sql window/browser tool.
if folks @ google listening (and know are), consider benefits of white-labeling browser tool: think you'd find lot of software companies integrating suites of products and, then, running circles around hadoop/cdh/emr/impala/hive combination.
so, summarize: how software developer import or emulate bq browser tool (with it's autocompletes, query histories, etc..) in own web-based app?
the initial version of bigquery web interface considered 'example' ui create themselves. uses public bigquery api talk bigquery.
there couple of google-internal things we've added since then, such current design of 'saved queries', , auth shortcut users don't have explicitly grant permission ui access bigquery data. still plain-ol-javascript talking bigquery via rest api same way else does.
the javascript obfuscated, however, understanding compression purposes downloads more quickly.
the sql highlighting done codemirror special configuration bigquery sql variant.
i'll talk other members of bigquery team open-sourcing javascript code in web ui. may difficult @ point, doesn't hurt have conversation it. i'll bring team , update thread. answer "we'll think it", can think , start working on :-)
let me know if sounds meet needs. might not solve auth problems mention, since users won't have bigquery accounts, may able solve proxying oauth2 access tokens.
Comments
Post a Comment