#725 and #761 are different. #761 and #714 are browser based (no need for server), they use the results that are already embedded in the page.
#725 is server based (Can handle full query result and bigger data) and do not use javascript. The interpreter is limiting the number of results sent to the UI so it doesn't explode, so both implementation are needed, your implementation could probably help the server side download (#725) On Thu, Apr 14, 2016 at 1:21 AM, Rusty Phillips <rusty.phill...@gmail.com> wrote: > I have seen a few questions on the mailing list related to accessing > resource pools from the front end, and some pull requests related to > downloading the result. > > (https://github.com/apache/incubator-zeppelin/pull/761, > https://github.com/apache/incubator-zeppelin/pull/725) > > Unfortunately, from brief inspection, these are working in javascript, so > cannot deal with data larger than the browser can handle (which is quite > easy to create with some of these big data tools). > > Fortunately, there is a design that addresses this (in part): > https://cwiki.apache.org/confluence/display/ZEPPELIN/Helium+proposal > > Wouldn't it be nice to use that same idea to allow for downloading TSV > files some day? Maybe today can be that day. > > Here are three tickets that cover using resource pools to download results: > > https://github.com/apache/incubator-zeppelin/pull/782 > https://github.com/apache/incubator-zeppelin/pull/816 > https://github.com/apache/incubator-zeppelin/pull/822 > > The third one contains everything that the first two do, but they're > probably worth reviewing independently because they're separate pieces of > functionality. > > What do you think? > > I bring it up on the mailing list because other pull requests quasi-related > to this are popping up, and I hope this can be merged so that those other > pull requests can make use of this to their advantage. >
