Alberto Islas

Thank you! working on these changes. About the `pacer_doc_id` question: Yeah, I've confirmed the four-digit `pacer_doc_id` is normalized to zero: [https://github.com/freelawproject/juriscraper/blob/main/juriscraper/pacer/utils.py#L82](https://github.com/freelawproject/juriscraper/blob/main/juriscraper/pacer/utils.py#L82) Here we’re also using the four-digit normalized to zero...

Perfect, thanks. So to avoid normalizing regular numbering I think we should only normalize document numbers from those courts that we know use the `pacer_doc_id` as document number, does it...

A small refactor was needed here. According to our talk about getting the document number for NDAs, the best approach is to get it before adding the docket entry/recap document....

@mlissner I've checked this error in detail and seems to be a bit more complicated than we thought. There are different potential errors that might need to be fixed, which...

Great, thank you for your answers! Now I think I have enough info that confirms the issue so I'll work to apply the solution to avoid the race condition that...

Perfect, well I'll start checking those other issues with the data we have now and see how they are related.

I've done some tests about this issue. I managed to reproduce a race condition that duplicates docket entries and recap documents. So I could assess some possible solutions. I think...

Of course, I checked this: To add a constraint to the DB seems that the whole table will be scanned: [ADD table_constraint](https://www.postgresql.org/docs/11/sql-altertable.html) > Normally, this form will cause a scan...

Thanks! I looked at some of these results and I found some that might indicate that sometimes these are not duplicates: For instance, the ones related to this case: [https://www.courtlistener.com/docket/42748/toru-paul-may/](https://www.courtlistener.com/docket/42748/toru-paul-may/)...

Of course. I only checked `select_for_update` independently (that didn't work) but I haven't checked it in combination with `selec_related`. I will give it a look and see if it works....