Mongodb large document design - mongodb

Question

I plan to create a database for price history.
The history database should store prices defined 90 days in advance each day in a year.
That means: 90 days x 365 days/year = 32850 database item
Is there any way to design schema to improve query performance ?
my first suggestion was hierarchical store values like:
{
"Address": "xxxxx",
"City": "xxxxx",
"Country": "Deutschland",
"Currency": "EUR",
"Item_Name": "xxxxxx",
"Location": [
log, lat
],
"Postal_code": "xxxx",
"Price_History": [
2014 : [
"January" : {
"CW_1" : { 1: [ price1 .. price90 ], 2: [ price1 .. price90 ], },
"CW_2" : {},
"CW_3" : {},
} ,
"February" : {},
"March" : {},
]
]
}
Thank you in advance!

Answer 1

It all depends on which queries you are planning to run against this data. It seems to me that if you are interested in keeping a history of actions, then your queries will almost always contain a date parameter.
The Price_History array might be better formatted as sub document. Each of these documents would have a varied (but limited) range of values - the year and the month. It might be a good idea to add an index on that attribute. This way, whenever you query by a certain date range, your indexes will assist mongo to find the relevant dataset relatively quickly.
Another option would be to have each price in-itself as a document. The item connected to the price could be a sub-document perhaps not containing all of the item data, but enough to be able to make the calculations and fetch the other relevant data once your dataset is small enough. For this usage, I would recommend creating a single attribute of the date ranges to be indexed and also an index on the item._id attribute. You can still have the individual date components if you still need to query them individually. Something like this:
{
"ind_attr": "2014_January_CW1",
"date": {
"year": 2014,
"month": January",
},
"CW": 1,
"price": [ price1... price90 ],
"item": {
"name": ...,
"_id": ...,
// minimal data about the actual item
}
}
With this document structure, you could easily add an index on the ind_attr attribute. The document.item._id attribute can be used to retrieve more detailed data on the actual item if needed.