Query specific ID's in a firebase year/month/day nest - firebase

Question

I have some data structured like the following in firebase. My client is built with angularjs + angularfire.
'all_records' is nested like this to split the data up and also as the majority of the querying is time orientated.
The 'all_records' nest is equal to: all_records > year > month > date > records.
I want to associate the records with the user that posted the record and therefore on record creation have saved the associated ID to the user in 'my_records'.
This example is obviously a very small dataset however to allow for scalability and potentially having having hundreds of thousands of records what would be the best way to query 'all_records' to get just the users 'my_records'.
Thanks
all_records : {
2015 : {
2 : {
3: {
-JgJJaLaL7LuuRL_IO7c : {
data: big data;
},
-JgJfaOsyfZe_mZPTApA : {
data: lots of data;
},
...
},
...
},
1 : {
28 : {
-JgJJaLaL7kddEl_ddfE : {
data: some more data;
}
},
27 : {
-JgJfadkdjEk_mZPTApA : {
data: even more;
}
},
...
},
...
},
2014 : {
12 : {
24: {
-JgJmnvkFdqc_y63lqcv : {
data: decemeber data;
},
-JgK6xbXY_E58lshaiVa : {
data: and a bit more data;
}
},
...
},
11 : {
...
},
...
}
}
user: {
name : James,
my_records : {
-JgK6xbXY_E58lshaiVa: true,
-JgJJaLaL7LuuRL_IO7c: true
}
}

Answer 1

Your user's my_records currently doesn't contain enough information to to directly map to the record. You'd need to add the date (year, month and day) to the information to allow a direct lookup.
I would instead structure things differently:
a list of just the records with their push-generated -J key
a list of users, with references to their records
a list/tree of dates, with references to the records for each date
Your application should never try to load the list of all records (#1). Instead it should load either the relevant user (from #2) or the relevant date(s) from (#3) and then load only the records that are referenced from there.
all_records : {
-JgJJaLaL7LuuRL_IO7c : { data: big data; },
-JgJfaOsyfZe_mZPTApA : { data: lots of data; },
-JgJJaLaL7kddEl_ddfE : { data: some more data; }
-JgJfadkdjEk_mZPTApA : { data: even more; }
}
dates: {
2015 : {
2 : {
3: {
-JgJJaLaL7LuuRL_IO7c : true,
-JgJfaOsyfZe_mZPTApA : true,
...
},
...
},
1 : {
28 : {
-JgJJaLaL7kddEl_ddfE : true
},
27 : {
-JgJfadkdjEk_mZPTApA : true
},
...
},
...
}
}
user: {
name : James,
my_records : {
-JgK6xbXY_E58lshaiVa: true,
-JgJJaLaL7LuuRL_IO7c: true
}
}
You might also want to consider structuring the dates as a simple list, instead of a tree:
dates: {
20150203: {
-JgJJaLaL7LuuRL_IO7c : true,
-JgJfaOsyfZe_mZPTApA : true,
...
},
20150128 : {
-JgJJaLaL7kddEl_ddfE : true
},
20150127 : {
-JgJfadkdjEk_mZPTApA : true
},
}
While it's tempting to build a tree for the dates, I doubt you'll get much benefit from it. The above structure is simpler and allow for a wider variety of queries, such as "give me all dates in Q1 of 2015": ref.child('dates').startAt('20150101').endAt('20150331').