Closed BilalBAI closed 3 years ago
It looks like this after removing that ValueError:
This will change the API / expected behaviour in a way that I am not comfortable with therefore I don't want to introduce this change. Note that a variant of this type of behaviour is available in a new package I released which supercedes pdblp
. You can take a look at specifying this type of behaviour in the new package here https://matthewgilbert.github.io/blp/quickstart.html#Troubleshooting
This will change the API / expected behaviour in a way that I am not comfortable with therefore I don't want to introduce this change. Note that a variant of this type of behaviour is available in a new package I released which supercedes
pdblp
. You can take a look at specifying this type of behaviour in the new package here https://matthewgilbert.github.io/blp/quickstart.html#Troubleshooting
Understood. Thanks @matthewgilbert .
When we use pdblp to request data for a list of tickers, if any one of the tickers is invalid, we get errors and can't get any data for the rest of the them. We have to pre-check all tickers one by one to handle the errors. It's very inefficient. If we remove this ValueError, it will return nan for invalid tickers which is easy to handle in the dataframe.
For pull requests please include the following:
git diff upstream/master -u -- "*.py" | flake8 --diff
Notes:
To run the
git diff
on theupstream/master
branch make sure this is set up, e.g.See here for more details.
Unfortunately automated testing using TravisCI is not possible since many tests require a valid Bloomberg connection. Please verify that
pytest pdblp/tests -v
has been run and all tests are passing.Commit
For most simple feature implementations please squash your commit to one entry and follow the style of https://chris.beams.io/posts/git-commit/ in your commit message, i.e.