I have been using CouchDB (trunk) for a couple of weeks now and while I
have a good grasp on the way it handles data I am unsure the best way to
store some of this data for retrieval purposes. I'll be honest upfront I
have some from an SQL background and so I'm still getting my head around
some of the concepts and while I could have done this in SQL without a
worry, I like the idea of the object and schema-less storage and the
replication option that CouchDB has.
Currently one of the document types I am storing is some time tracking
data it look a little like this:
{
"|_id|":|"t125142312603660"|,
"|_rev|":|"1-14095bf3c015575a4dc5ec3c7aea1234"|,
"|task|":|"blah"|,
"|cc|":|"HYD"|,
"|timestamp|":|1251423124|,
"|pc|":|"SPR"|,
"|duration|":|5|,
"|username|":|"andrew"|
}
The timestamp field is a utc unix timestamp.
I have been using views with map and reduce functionality to collate
this data into the data I want.
However my issue comes when I want to only look at data between
particular time stamps.
I thought about collating it for a given day do it looked like (format
is bad I know!):
key:["2009-08-26","HYD","SPR","blah"], value: 100 ;
where value is summed value of durations. But this doesn't work for
situations of different timezones.
I thought that I may be able to use startkey/endkey docid and use the
time stamp as the docid to filter the documents used for the reduce -
but this doesn't seem to be the way it works. (Am I wrong??)
I did think about using an external process, but this only works as both
a map and reduce not just a reduce option.
I am interested to know your thought on the best way to handle this data
and retrieve it?
Regards
Andrew M
