Re: Strange results with date_trunc 'QUARTER'

Looks like a bug to me. Please open a ticket. A simple repro would be very useful. https://issues.apache.org/jira/browse/DRILL - Sudheesh On Apr 3, 2017, at 2:11 PM, Joel Wilsson mailto:joel.wils...@gmail.com>> wrote: Hi, I'm seeing some strange results when trying to group by date_trunc('QU

Re: S3 using IAM roles

Glad you could figure this out! Can you open a ticket with details? https://issues.apache.org/jira/browse/DRILL - Sudheesh On Apr 3, 2017, at 12:42 PM, Knapp, Michael mailto:michael.kn...@capitalone.com>> wrote: All, In case others are in the same situation as I am, I will tell you how I solv

Re: Drill Parquet Partitioning Method

I filed a JIRA for this a while back: https://issues.apache.org/jira/browse/DRILL-4615 And even more recently finished a patch on our internal branch (pegged at 1.6 right now). Since 1.6, things have shifted a bit in the abstraction of managing directory partitions, so the patch wouldn't be direct

Strange results with date_trunc 'QUARTER'

Hi, I'm seeing some strange results when trying to group by date_trunc('QUARTER', ). I can work around it by doing more or less the same thing as in DateTruncFunctions. Am I missing something, or is this a bug? 0: jdbc:drill:> SELECT date_trunc('QUARTER', `taxi_trips`.`dropoff_datetime`), COUNT(*

Re: Drill Parquet Partitioning Method

That's a good idea. Let me clarify one thing first. Drill has two kinds of partitions: auto partition, or directory-based partition. The first one is a result of using drill's CTAS partition by statement [1]. Both partition column name and column value are written and encoded in the output file.

Re: S3 using IAM roles

All, In case others are in the same situation as I am, I will tell you how I solved this. After A LLL of digging through source code, I discovered the following facts: • Drill is using hadoop’s FileSystem to support S3 queries. So any configuration items that work for that will also w

Drill Parquet Partitioning Method

So as a user of Drill now for a while, I have gotten used to the idea of partitions just being values, instead of key=value like other things (hive, impala, others). >From a user/analyst perspective, the dir0, dir1, dirN methodology provides quite a bit of flexibility, but to be intuitive, we have

S3 using IAM roles

Drill Developers, I am using IAM roles on EC2 instances, your documentation here: https://drill.apache.org/docs/s3-storage-plugin/ instructs me to provide an access key and secret key, which I do not have since I am using IAM roles. I have been reviewing the source code a few hours now and stil

Re: Reading Drill generated Timestamp from spark

Hi Rahul, According to the parquet specification the primitive datatype for the TIMESTAMP logical type is INT64 (INT96 has a deprecated label). That's why Drill has not a mechanism to generate such values. But