16 comments

[ 3.3 ms ] story [ 50.5 ms ] thread
I couldn't find any information on persistence and crash recovery for this thing... is it able to beat Volt because it doesn't write to disk?
In fact we do write a redo log to disk (Set the REDO environment variable, e.g. REDO=$filename bin/driver when running the driver in our example download). VoltDB tries to be fast at highly partitionable workloads and does not provide OLAP-style query support. HyPer aims at bringing OLTP and OLAP together into one system.
Peter Boncz is involved in some capacity, he told me last time we talked. I believe Peter is the senior academic from the VectorWise project, and second most senior generally in the MonetDB gang after Martin Kersten.
Stumbled upon it a few months ago, but nowhere to download it unfortunately. For OLAP needs, the best solution as far as I know right now is still KDB+ ( managed to make some complicated queries with aggregates on 2 millions rows in 46ms on a macboook air with it, quite amazing )... Any alternative you know ?
KDB+ .. looks to have its own language to query it .. btw you can download hyper-db from http://www.hyper-db.com/downloads/hyperdemo.tar.xz
Yes, that is our official download link which contains a driver to run transaction benchmarks (specified in our own script language; the tar contains an example for TPC-C) and a psql-like sql tool with which you can load data and query it using SQL-92. An example for TPC-H is also included. You can contact us/me for further questions (muehlbau@in.tum.de).
I have 3 Questions --

1) How you have created the web interface as I can't see any php or python bindings.

2) How do you copy the csv data into the memory table

3) Can we use it internally as our company's internal BI database , since no licenses are mentioned in the downloaded link so its not very clear ..

1) The component used for the web interface is not part of the demo

2) Check out hyperdemo/scripts/tpch/load.sql, HyPer uses the same syntax as Postgres, i.e. you run bin/sql, create a customer table and load it with

  copy customer from '../tpch/customer.csv' delimiter ',';
3) Please contact us http://www.hyper-db.com/#team
> KDB+ .. looks to have its own language to query it

And that's a good thing.

You can use sql to query it if you want by starting the query with "s)" in the console, or simply by using the ODBC/JDBC drivers.

But KDB+ embraces order in a way that SQL does not, which means that if you limit yourself to SQL then you miss out on a lot power that KDB gives you.

For example, say you want to select a list of (date,teacher) records ordered by date, and out of these pick out the records in which the teacher is not the same as in the previous date.

In kdb+, it's "select from (select from records by date) where differ teacher" (or just "&~=':teacher[<date]" if you use K notation). What's the SQL equivalent?

Any more In Memory Databases ..which you have tried ?