On Thu, Apr 15, 2010 at 5:45 PM, Pierre-Alexandre St-Jean <p...@stjean.me>wrote:
>
> RowKey=time (or sequence id)
>> Usually this is non-optimal.
>>
>> If your keys are in order, inserts only go to one region and you do not
>> get
>> the throughput from multiple regions. Since rowkeys always have an
>> implicit
>> timestamp your rowkey should probably be something else.
>>
>
> So I got multiple variable that i want to keep the value they had over
> time.
>
> I get the row key to be the variable name.
>
> My limit is the maximum number of versions a column/rowkey can have(or is
> it a bug when you create a family in the shell). ~1536000000
>
> I will think of the optimal way to do it. If someone can help me
> --
> Pierre-Alexandre St-Jean
>
This is my modelling after json
Point{
name:"pointname",
attributes:{
description:"The point description",
unit:"the point mesurement unit"
}
values:[
{
timestamp:ts1,
value:"pointAtTs1"
},
{
timestamp:ts2,
value:"pointAtTs2"
}
]
}
This is what i think i will do : This way i think it will permit analysis
based on time and point name, and permit modification of values if there has
been a bad entry (but then there was a problem ( but that is another
thing)). I dont know if i will do it in one or 2 tables.
rowkey:"pointname"{
attributes:{
description:{"description",ts1}
unit:{"unit",ts1}
}
}
rowkey2:"pointname/logginTimeStamp"{
value:{"value at logginTimeStamp",ts1}
}
