Generally, data cubes are designed to never delete data. Our reports
are accumulative so it's not an everyday issue. We tend to reuld the
data if there is a problem and, as LOAD DATA is so quick and the
queries themselves so potentially time consuming, this is not
perceived as much of an inconvenience.
You are not going to find full-text search or an other indexing
structures in this storage engine.. Those operations are always
expensive and therefore don't belong in the data warehousing space.
The idea is to boil down your raw data into pure integer structures
before it gets into the warehouse.. that's when reporting on massive
amounts of data becomes feasibile. Any verbal content gets reduced to
indexes into a global dictionary.
It's important to note that data warehousing, business intelligence,
data cubing, whatever you want to call it, is fundamentally different
than the operational business data which is the backbone of our
- michael dykman
On Sun, Feb 8, 2009 at 5:51 PM, mos <mos99@stripped> wrote:
> At 11:14 AM 2/8/2009, Michael Dykman wrote:
>> We recently began using infobright in my shop.. reporting queries
>> against 10's of millions of rows which previously had taken ~2 hours
>> to execute where reduced to less than 10 minutes. It certainly comes
>> with it's own set of limitations but it has been a god-send for us.
>> - michael
> Michael & Baron,
> I took a quick look at InfoBright and it seems interesting but...
> 1) Infobright appears to be a read-only database engine with the only means
> to update data is through "Load Data Infile" and Insert, Updated, Delete are
> not supported? Is that correct? How do you get rid of old rows or update
> existing rows?
> 2) Does it have Full-Text search? I need to search on paragraphs of text.
>> On Sun, Feb 8, 2009 at 10:58 AM, Baron Schwartz <baron@stripped> wrote:
>> > I'd not hold your breath.
>> > Take a look at the open-source version of Infobright.
>> > On Sat, Feb 7, 2009 at 8:21 PM, mos <mos99@stripped> wrote:
>> >> NitroEDB was suppose to be shipped Q2 2007. Whatever happened to that?
>> >> When
>> >> can I expect it? TIA
>> > --
>> > MySQL General Mailing List
>> > For list archives: http://lists.mysql.com/mysql
>> > To unsubscribe: http://lists.mysql.com/mysql?unsub=1
>> - michael dykman
>> - mdykman@stripped
>> - All models are wrong. Some models are useful.
> MySQL General Mailing List
> For list archives: http://lists.mysql.com/mysql
> To unsubscribe: http://lists.mysql.com/mysql?unsub=1
- michael dykman
- All models are wrong. Some models are useful.