kudu-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dan Burkert <danburk...@apache.org>
Subject Re: KuduScanToken pushdown question
Date Fri, 21 Jul 2017 16:57:45 GMT
Hi Clifford,

Currently there isn't a way to do that.  If you are 100% sure the PK ranges
don't overlap, you might consider creating multiple sets of scan tokens,
each with a unique range (through separate ScanTokenBuilder instances).
This is more or less what Kudu would do behind the scenes to support
disjoint ranges, anyway.

- Dan

On Fri, Jul 21, 2017 at 8:43 AM, Clifford Resnick <cresnick@mediamath.com>

> I’m playing with Drill’s kudu-storage code and converting it to use the
> ScanToken api. It’s a fairly simple matter to do this by serializing
> ScanTokens (thanks of the great api!), but I’m not sure how much to “push
> down” to Kudu. For Tablet/Drillbit affinity and for pruning it’s clear I
> need to push hash key and range bounds predicates, but it can get a bit
> tricky given that the ScanToken api does not support OR logic unless it
> fits into an IN_LIST expression, so for more complex logic I suppose it’s
> best to not push down the filtering.  However, I’m wondering if there is a
> way to push disjoint bounds to Kudu. For example, if I have a table with
> range keys of Year,Month,Day, it there a way I can include only (2017,6,1)
> OR (2017,5,5) in a group scan?
> Thanks,
> Cliff

View raw message