Tml 2022 backend 1st column id mark as primary key #3683
Conversation
Codecov Report
@@ Coverage Diff @@
## main #3683 +/- ##
=======================================
+ Coverage 99.7% 99.7% +0.1%
=======================================
Files 337 337
Lines 34067 34077 +10
=======================================
+ Hits 33936 33946 +10
Misses 131 131
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
…github.com:alteryx/evalml into TML-2022-backend-1st-column-id-mark-as-primary-key
| check_all_unique = X.nunique() == len(X) | ||
| # Temporary solution for baton logical types mapping integers to doubles in woodwork logical types. | ||
| # Will be removed when resolved. | ||
| check_all_unique = X_double.nunique() == len(X_double) |
There was a problem hiding this comment.
To reduce the re-used code, I wonder if we can do either a for loop:
for dtypes in [['Double'], [Integer', 'Categorical']]:
# logic hereor to separate out this into a helper method, where we can pass in the Double and Integer/Categorical arguments. I know this is just a stopgap fix for now as we wait, but might be nice to not have repetitive code and reuse parts when we can.
| check_all_unique | ||
| ].index.tolist() # columns whose values are all unique and doubles | ||
| cols_with_all_unique_integers = [ | ||
| col for col in cols_with_all_unique if all(X_double[col].mod(1).eq(0)) |
There was a problem hiding this comment.
I wonder if we can use
col for col in cols_with_all_unique if all(X_double[col].is_integer())to capture this logic
There was a problem hiding this comment.
This doesn't work unfortunately because .is_integer doesn't work for series types
| @@ -1,4 +1,6 @@ | |||
| """Data check that checks if any of the features are likely to be ID columns.""" | |||
| from xml.etree.ElementInclude import include | |||
There was a problem hiding this comment.
not sure where that came from but I removed it!
simha104
deleted the
TML-2022-backend-1st-column-id-mark-as-primary-key
branch
Pull Request Description
IDColumnCheck now handles primary key columns containing "integer" values that are typed as doubles