Filter query

[harshitsinghai77]

This is how my deta looks like

I want to query all the entries that have a created_at in the last 7 days. But my queries return an empty list.

query = {"key": user_id, "analytics.duration?lt": str(datetime.now())}

I even tried, this is also empty,

query = {"key": user_id, "analytics.duration": 4501}

what am I doing wrong?

[jnsougata]

analytics is a list so you have to do analytics[0].duration?lt: str(datetime.now())

The way you are storing the data is not much efficient to make an exact query.
Rather store the data with individual user/keys. It will make extraction easy and faster and more efficient.
You can see table below for reference:

and there are two attributes to the user anlytics clicks and retention and stored as the table below:

and the query for all the users having clicks > 800 will be {'analytics.clicks?gt': 800}
and you will receive a response as:
{'paging': {'size': 2}, 'items': [{'analytics': {'clicks': 900, 'retention': 0.15}, 'key': 'user_8'}, {'analytics': {'clicks': 1000, 'retention': 0.16666666666666666}, 'key': 'user_9'}]}

[abdelhai]

i hate to say it but you are using base wrong. for queries to work and work efficiently you need to plan out your schemas etc.
i don't recommend putting all user events in the user object. the correct approach is to use a separate base for events and another one for users.

[jnsougata]

If separate base is assigned for each individual user it will be hard to implement caching locally. So I prefer the single base pattern where I can cache the whole base at start up and update in realtime. In this way I can reduce API calls as well.

[abdelhai]

i didn't mean a base for each user, rather one base for all events and one base for all users.

[jnsougata]

oh mb

[harshitsinghai77]

Got it. Thanks for the help.