Add ability to ignore invalid ids #23
Conversation
|
@OliverSherouse have you had a chance to review? |
|
I know it's been a good while, but can I ask for you to rebase this off the current branch? If you're no longer interested just let me know and I'll integrate these changes myself. |
|
Yeah no problem, I can rebase. Thanks for getting back on this. |
Added a new keyword argument for get_series which allows for invalid ids to simply be ignored rather then fail. This will avoid unecessary requests for series that were valid, but happened to be grouped with a invalid series. Had to also rewrite a few of the tests since this new version returns a dataframe instead of series when one one series is requests.
|
@OliverSherouse, I resolved all conflicts and re-pushed. Checks are failing though, but I believe it is unrelated. Either way, let me know if you have any thoughts or questions. |
|
Thanks! I'm also a bit unclear on what's happening with that test, so I'm going to redirect into an |
* Add ability to ignore invalid ids (#23) Added a new keyword argument for get_series which allows for invalid ids to simply be ignored rather then fail. This will avoid unnecessary requests for series that were valid, but happened to be grouped with a invalid series. Fix some typing issues. Closes #30 Co-authored-by: Maxim Kupfer <mbkupfer@users.noreply.github.com>
@OliverSherouse, I think it would be a vast improvement if we can add the option to ignore invalid ids. There are times where I have just one or two invalid series ids in the mix, but the rest are fine. I don't see why the whole get_series request should fail because of those couple wrong ids.
The new approach still let's the user know if there was a bad id (or ids), but it instead throws a warning message instead. This all requires using a new keyword argument though, called errors. I modeled this off of pandas.Dataframe.drop which also has the same type of argument.
Besides that change, I also had to adjust a few of the test cases since a dataframe is now being returned instead of a series. Reason is because I'm now using an array to create the final returned dataframe and that coerces the series into a dataframe. In the end, I actually think this is an improvement since the datatype doesn't change from the enduser perspective.