Conversation
|
|
||
| date_obj = datetime.datetime.strptime(raw, DATE_FORMAT) | ||
| date_obj = date_obj - datetime.timedelta(days=1) | ||
| date_obj = date_obj - datetime.timedelta(hours=6) |
There was a problem hiding this comment.
I am reluctant to make this "default" change. If we need to do this, I think the best course of action is to make this a sort of config change where users can pass in their desired time and the tap adjusts accordingly.
|
@cmcarthur Do you know why this was set to 1 day? |
|
@amyfain Would it be possible to lower the replication frequency- like run the tap on 3 hour intervals- to combat the duplicate data issue? |
|
@luandy64 while that's possible, we have our replication frequency set to 1 hour because we want the data as quickly as possible. I'm not sure how to make this a config change, but I think that's a great idea! |
|
@amyfain We can iterate on this idea- and I invite you to check out the Singer Slack if you aren't there already- but I think it could go like:
There might be a better way of doing this, but I'll sleep on that 👍 |
This reduces the bookmark rewind. When it is set at 24 hours and we run the tap in 1 hour intervals, we see the same data pass to us 24 times as it's set right now.