Best way to pull one random record from Deta Base? (Python)

[eddy-22]

Hi all,

I have a LARGE Base and was wondering what the best way to get one single record, at random, would be. Something similar to MongoDB's aggregation:

db.collection.aggregate( [  { "$sample": {
  "size": 1
} } ] )

Or something similar to pymongo's count_documents and find_one functions:

count = collection.count_documents( {} )   # using an empty filter
item = collection.find_one( {} , skip=count) 

---

I have thought of two ways to do this, which would be the best route, or is there another way that would be more performant? My thoughts were:

1. Fetching all the items using the example in the documentation, then taking a random record from that list:

res = db.fetch()
all_items = res.items

while res.last:
  res = db.fetch(last=res.last)
  all_items += res.items

import random
desired_item = random.choice(all_items)

However this is going to be very time consuming and inefficient due to my base's size.

2. Use db.get(key) with a range of incremented keys, and taking a random key from that to pass to db.get. I'm not sure how to implement this at the moment.

---

If anyone has suggestions or better ideas please let me know, thanks!

[jnsougata]

First I want to know what type of keys you are using to store the records?
Is those maintaining a sequence or any data field inside your base is having a data maintaining a sequence for all the keys ?

[eddy-22]

Yes, so I actually have a timestamp field that is sequenced ascending. The key fields are currently strings of n, n+1, n+2 etc. The first record key="1", the next="2", the next="3" and so on (as strings).

[jnsougata]

So if you know the range of the key then you can just generate a random integer randint() then convert it to str and use .get() to fetch the record from base (assuming the range will be fixed otherwise you have to check and retry)

key = str(random.randint(first, last+1))
db.get(key)

[abdelhai]

deta base doesn't support this capability. the randomization should happen outside the query. in addition to what @jnsougata suggested, it might also be possible to generate a specific time in the past and use comparison operators to get a random result. not sure how well this would work in reality.