Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Oct 20 18:08
    zmz0305 synchronize #986
  • Oct 20 17:52
    zmz0305 synchronize #986
  • Oct 18 16:34
    zmz0305 synchronize #986
  • Oct 18 16:00
    zmz0305 synchronize #986
  • Oct 18 15:43
    zmz0305 synchronize #986
  • Oct 17 21:22
    zmz0305 synchronize #986
  • Oct 17 18:42
    zmz0305 synchronize #986
  • Oct 17 05:11
    zmz0305 synchronize #986
  • Oct 17 05:03
    zmz0305 opened #986
  • Oct 15 20:00
    michael-mclawhorn synchronize #982
  • Oct 11 18:46
    michael-mclawhorn opened #985
  • Oct 11 18:45
    michael-mclawhorn closed #977
  • Oct 11 18:45
    michael-mclawhorn closed #979
  • Oct 11 18:40
    michael-mclawhorn closed #984
  • Oct 09 17:23
    efronbs closed #981
  • Oct 09 16:23
    michael-mclawhorn opened #984
  • Oct 09 15:39
    efronbs synchronize #981
  • Oct 08 22:01
    archolewa synchronize #982
  • Oct 08 18:29
    efronbs synchronize #981
  • Oct 08 17:17
    efronbs synchronize #981
Andrew Cholewa
@archolewa
Though someone needs to plus one it, and we need to apparently kick off the build again, because it's failing, despite being all text...
I've rebased #442 on top of master.
Kevin Hinterlong
@kevinhinterlong
Ok. Do we merge RFC PRs?
I can still give it a :+1: but the only one I’ve seen before is titled “RFC and design for composite table availability design do not merge"
Andrew Cholewa
@archolewa
I think we should. Having documentation describing the design and motivation of a feature is quite valuable, in my opinion. It can serve as the seed for documentation describing the feature, and it can give people context for understanding the feature.
Kevin Hinterlong
@kevinhinterlong
Agreed :smile:
+1d
Kevin Hinterlong
@kevinhinterlong
Thoughts on a self-reporting fili module?
We could log some metrics into a sql database and expose it as a logical table using fili-sql
Kevin Hinterlong
@kevinhinterlong
@michael-mclawhorn rxjava2 built + passed tests https://travis-ci.org/yahoo/fili/builds/263576431
Andrew Cholewa
@archolewa
@michael-mclawhorn I understand Tarrant's problem now. Our Physical tables imply really really really strongly that the physical table name and the data source name are not necessarily the same. However, the VolatileDataRequestHandler
assumes that they are the same.
Andrew Cholewa
@archolewa
Unfortunately, this is going to be tricky to fix, because in a world where physical tables can be views on a data source, the VolatileDataRequestHandler doesn't have enough information to fulfill its function. Once we've built the Druid query, we don't know for sure which physical table was actually selected.
*Tricky to fix for every case. You can get it to a point now where it'll work so long as you don't have different physical tables with different volatilities backed by the same data source.
tarrantzhang
@tarrantzhang
Hi all! Is there a reason why TableDataSource constructor takes in a ConstrainedTable rather than a general PhysicalTable?
Kevin Hinterlong
@kevinhinterlong
Maybe you could use physicalTable.withConstraint(DataSourceConstraint.unconstrained(physicalTable))?
Kevin Hinterlong
@kevinhinterlong
@tarrantzhang ?
tarrantzhang
@tarrantzhang
@kevinhinterlong Thanks, this should deal with my problem. Still, I'm curious about the reason behind why TableDataSource takes a constrained version of table.
Kevin Hinterlong
@kevinhinterlong
My guess from just looking around would be something to do with serialization and that being used for building druid queries
tarrantzhang
@tarrantzhang
You are right. I traced down to the source and TableDataSource is used in druid query builder which used a ConstrainedTable.
michael-mclawhorn
@michael-mclawhorn
yahoo/fili#519 @QubitPi @garyluoex
Andrew Cholewa
@archolewa
@michael-mclawhorn Could you take a look at this PR in the next few days? https://github.com/yahoo/fili/pull/489/files It's a PR by @mpardesh to remove the ui/non-ui Druid broker. @kevinhinterlong has already approved it. I know that this is a thing you've wanted simplified for a while, so I figured you might want to take a look. @mpardesh is going back to school this weekend
though, so I'd like to get her stuff closed out by then.
yahoo/fili#502 could also use a look. It adds "first" and "last" as pagination parameters. @kevinhinterlong and I have both already gone through it once, and I'm going to go through it a second time this morning.
Andrew Cholewa
@archolewa
@QubitPi Could you rereview this PR? yahoo/fili#429 You requested some changes on it. Since it already has two approvals, and I don't want it to sit in perpetuity, I'm going to dismiss your review and merge it in tomorrow morning if you don't get a chance to look at it by then.
Note that your requested changes have been addressed.
Kevin Hinterlong
@kevinhinterlong
@archolewa I updated my review on yahoo/fili#502
Andrew Cholewa
@archolewa
@kevinhinterlong Thanks!
Now go back to enjoying what summer vacation you have left.
Andrew Cholewa
@archolewa
@here yahoo/fili#523 Replaces yahoo/fili#429 with a copy of the branch on the main Fili fork, so that we can continue to keep it up to date when @mpardesh goes back to school, until we can finish creating our docker hub. If I could get a few people to take a look and approve it, that would be great.
Monica Pardeshi
@mpardesh
@michael-mclawhorn Do I need to make any changes to yahoo/fili#507 ?
Andrew Cholewa
@archolewa
@mpardesh It looks like we're getting a compilation error for yahoo/fili#489
Yongli Chen
@saiyan86
@tarrantzhang I just updated yahoo/fili#499 could you please give it a look?
@QubitPi
Dejan Stojadinović
@dejan2609
@will-lauer please review: yahoo/fili#459
michael-mclawhorn
@michael-mclawhorn
yahoo/fili#542
michael-mclawhorn
@michael-mclawhorn
yahoo/fili#543
michael-mclawhorn
@michael-mclawhorn
yahoo/fili#544
kiran sunkari
@kiransunkari
hi anyone there..?
Rick Jensen
@cdeszaq
Hi @kiransunkari, what’s up?
Jiaqi Liu
@QubitPi
Review needed :wink: yahoo/fili#582
Rick Jensen
@cdeszaq
@QubitPi What's the story with yahoo/fili#697 ?
It looks like it shouldn't have been merged, but it was. Any objection to me rolling it back?
Jiaqi Liu
@QubitPi
Yeah I was also surprised to saw it had been merged before I got a chance to take a first glance at this PR :( Looks like @michael-mclawhorn had some concerns after it was merged as well. I don't mind rolling it back
It appears to me that this PR was for an urgent demand from Deepak's team. But yeah, I think at least one core team review is mandatory before that PR can merge
Jiaqi Liu
@QubitPi
@michael-mclawhorn @archolewa @cdeszaq I've addressed your comments for yahoo/fili#688 & yahoo/fili#690 . They are ready for review again
Jiaqi Liu
@QubitPi
@michael-mclawhorn @archolewa @cdeszaq Could you take a look at yahoo/fili#688 ?
Rick Jensen
@cdeszaq
Will do, likely tomorrow evening though
Jiaqi Liu
@QubitPi
Thank you, Rick!
Jiaqi Liu
@QubitPi
@cdeszaq yahoo/fili#688 is ready for review. Could you take a look at it?
Jiaqi Liu
@QubitPi
@cdeszaq yahoo/fili#748 needs some review if you have some time :)