Skip to content

Conversation

@jtratner
Copy link
Contributor

Via verify_integrity kwarg. Internals all pass False.I don't think I've missed anything / passed verify_integrity=True where
it ought not to be.

Fixes #5213.

Clearly this adds a (likely small) perf hit - but should only occur
from external MI calls that specifically call the constructor with
levels and labels (for example, doesn't occur with from_arrays or
from_tuples). For 0.14 I'll have to do it differently, but this should
work for 0.13.

@jtratner
Copy link
Contributor Author

Changed it to validate pickles too.

@jtratner
Copy link
Contributor Author

people okay with this? @wesm?

@jreback
Copy link
Contributor

jreback commented Oct 15, 2013

good by me

@wesm
Copy link
Member

wesm commented Oct 15, 2013

Performance impact probably minimal so this is fine to go in.

@jtratner
Copy link
Contributor Author

Okay I'll rebase. I might add something to the docs about how setting levels or labels triggers a new round of validation.

Via verify_integrity kwarg. Internals all pass False. CLN: Verify integrity by default on label/level setting
jtratner added a commit that referenced this pull request Oct 16, 2013
BUG/CLN: MI now checks level & label compatibility
@jtratner jtratner merged commit f70a6e2 into pandas-dev:master Oct 16, 2013
@jtratner jtratner deleted the mi-validate-level-label-compat branch October 16, 2013 02:07
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

3 participants