from Hacker News

Show HN: EventQL – Open-source distributed SQL analytics database in C++11

by paulasmuth on 7/26/16, 3:39 PM with 18 comments

  • by jimktrains2 on 7/26/16, 6:42 PM

    Is this similar to Apache Drill[0]/Google Dremel[1] (BigQuery[2])? One differences is that it seems to be able to do mutations to data, not simple appends.

    [0] https://drill.apache.org/ [1] http://research.google.com/pubs/pub36632.html [2] https://cloud.google.com/bigquery/

  • by babas on 7/26/16, 7:08 PM

    Looks great! I've just gotten a single instance up and running. Really simple to set up. It seems almost exactly what We've been looking for! Some background: We have been evaluating different time series databases for use with sensor readings (We probably need to ingest something on the order of 100,000-200,000 samples/s per cluster). The data is from physical sensors like temperature, level, pressure etc. Where all the data is in the form of tag(string)|datetime(ms)|value(decimal).

    Have you done any comparisons to other similar software like influxdb, Cassandra, etc? Especially ingestion rate and disk usage.

    What kind of pricing can we expect on Managed Hosting?

    We are currently leaning towards Influxdb but the cluster licensing stuff they are doing really made us think twice.

  • by dtheodor on 7/26/16, 9:21 PM

    Why EventQL over other distributed columnar-storage based databases like Redshift, Vertica, Citus, or Hadoop-based ones like Impala and Presto. What does EventQL do better?

  • by nwrk on 7/26/16, 4:54 PM

    looks good [0], thank you

    can anyone comment on data disk usage ?

    [0] docs - http://eventql.io/documentation/

  • by ddorian43 on 7/26/16, 5:45 PM

    Why not build a bigtable-cassandra fusion row-store too ? Since updates are async and all nodes are the same (cassandra) while the data model is sorted-by-primary-key (bigtable) and the schema fixed (low cost in storing tuples) and sql available (easier for devs).

  • by buremba on 7/26/16, 6:40 PM

    How can we import CSV dataset? HTTP API for insert returns this error message: expected JSON_OBJECT_BEGIN, got: JSON_OBJECT_BEGIN. The GROUP BY query in homepage scans 1.8B rows and only takes 1.5 seconds which is great but how many nodes used in that setup?

  • by ktamura on 7/26/16, 5:37 PM

    As a maintainer of the projec,t it's cool to see the Fluentd plugin in the repo: https://github.com/eventql/fluent-plugin-eventql