We recommend you use the following practices when requesting data, especially for larger brands:
- Consistently using a LIMIT of 100000
- No more than 12 concurrent requests
- If you're collecting data for a Brand for the first time, use the
eventDate
date parameters to retrieve all of the historical data, and once that's complete (there's no more historical data present), switch to using theaffluentModifiedAt
parameters moving forward. That will mean you are only pulling new actions, and also updates to any historical actions (see below for more details on why this is recommended).
Collecting historical data from a Brand for the first time/Collecting data from a recently connected account
- If the account has been recently connected to Affluent or you're collecting data for a Brand for the first time, you may notice that it's taking a very long time to pull actions using the
affluentModifiedAt
date parameters (you will only encounter this if it's a large enough brand). In these cases, consider using 1 day date ranges in your request.- If you’ve tried that and it still takes too long, consider using
eventDate
parameters instead, so that you can retrieve all of the historical data in smaller, more manageable chunks.- Once you’ve retrieved all of the historical data, go ahead and switch to using
affluentModifiedAt
date parameters so that you can pull in new actions and maintain your data store by updating any modified actions.
- Once you’ve retrieved all of the historical data, go ahead and switch to using
- For context: when you connect a brand, Affluent will pull in all of the historical data for that brand in a short time span, meaning all of the
affluentModifiedAt
times for all historical actions are going to be crammed into that short time span - that could be millions of rows per day. It’s more manageable in these cases to break down the data you request usingeventDate
parameters so that you’re able to send multiple smaller requests that perform a lot faster.
- If you’ve tried that and it still takes too long, consider using