RegisterSign In
By Rob Bygrave (rbygrave) on Aug 19, 2009 9:07 PM.
Integer, Double, Date, Timestamp data types

I'm wondering if you have plans for direct support for these types.

I am wondering about a typical OLTP app that frequently is using these types in predicates. That is, to get great performance on things like

amount > 10.50
orderDate = Date(2009-08-20)
lastUpdate > Timestamp(...)


Another way of asking the question would be... are you confident of getting good performance for these types of expressions without built in support for these types?
By Tobias Downer (toby) on Aug 19, 2009 9:52 PM.
SDBTable supports indexes but the only type it supports are strings (lexicographical order). You can encode the types you mentioned as strings, for example, numbers would need to be encoded as '0000000099', '0000000100', etc. Then you could do an index query of tail('0000000090') to find all rows > 90. Timestamps can be encoded similarly. BigTable works like this - they have a number of utility classes for encoded various types as strings.

You should get good performance (it's a tree search regardless of whether the db supports more complex types or not), and it scales horizontally over your cluster anyway.
By Tobias Downer (toby) on Aug 19, 2009 10:40 PM.
Oh also, yes, we do plan to release a data model with built in types for dates and numbers, etc.
Please sign in or register to post in this topic.
The text on this page is licensed under the Creative Commons Attribution 3.0 License. Java is a registered trademark of Oracle and/or its affiliates.
Mckoi is Copyright © 2000 - 2019 Diehl and Associates, Inc. All rights reserved.