-
-
Notifications
You must be signed in to change notification settings - Fork 25.8k
Make labelencoder use hashtable #7455
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make labelencoder use hashtable #7455
Conversation
Not sure why you closed, but we can't have pandas as a dependency. On 19 September 2016 at 17:06, themrmax notifications@github.com wrote:
|
@jnothman i closed because i noticed all the commits from my other unmerged pull request (can you take a look at that again btw? i fixed the issue with the sentinel.) how come we can't use pandas as a dependency? |
I'm not sure what that means.
Why should we? Minimising dependencies is generally a good idea to avoid issues with versioning, compatibility testing, installation overhead, etc... |
the pull request has 13 commits on it but i only want the last one. the reason i would use pandas is because |
Sometimes you have to compromise or reimplement. On 20 September 2016 at 10:09, themrmax notifications@github.com wrote:
|
Reference Issue
Fixes #7432
What does this implement/fix? Explain your changes.
Use a dict lookup instead of
np.sortedsearch
for the label lookups. Also removed the customfit_transform
method, since the dict lookup is quicker than thenp.unique(return_index=True)
trick that was used here before.Any other comments?