AutoMLSearch: data splitting fails with IndexError

Problem

Index Error when search is called

Reproducer:

Code

X = pd.read_csv('data.csv')
y = X.pop('TARGET')
automl = AutoMLSearch(problem_type='binary', objective='f1')
automl.search(X, y)

Data:

https://alteryx0-my.sharepoint.com/:u:/g/personal/ethan_tu_alteryx_com/EcgNVHHRUW1Gi9wcWXcOc3oBp_VDRhksoHopj5r8UWF1VA?e=vCCmhs

Stack Trace:

---

IndexError                                Traceback (most recent call last)
~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/indexing.py in _get_list_axis(self, key, axis)
   2110         try:
-> 2111             return self.obj._take_with_is_copy(key, axis=axis)
   2112         except IndexError:

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/generic.py in _take_with_is_copy(self, indices, axis, **kwargs)
   3408         """
-> 3409         result = self.take(indices=indices, axis=axis, **kwargs)
   3410         # Maybe set copy if we didn't actually change the index.

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/generic.py in take(self, indices, axis, is_copy, **kwargs)
   3393 
-> 3394         new_data = self._data.take(
   3395             indices, axis=self._get_block_manager_axis(axis), verify=True

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/internals/managers.py in take(self, indexer, axis, verify, convert)
   1385         if convert:
-> 1386             indexer = maybe_convert_indices(indexer, n)
   1387 

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/indexers.py in maybe_convert_indices(indices, n)
    212     if mask.any():
--> 213         raise IndexError("indices are out-of-bounds")
    214     return indices

IndexError: indices are out-of-bounds

During handling of the above exception, another exception occurred:

IndexError                                Traceback (most recent call last)
<ipython-input-18-4575b463e46a> in <module>
----> 1 automl.search(X_train, y_train)
      2 pipeline = automl.best_pipeline
      3 pipeline.fit(X_train, y_train)
      4 
      5 scores = evalml.pipelines.calculate_permutation_importance(pipeline,

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/evalml/automl/automl_search.py in search(self, X, y, data_checks, feature_types, raise_errors, show_iteration_plot)
    365 
    366         start = time.time()
--> 367         self._add_baseline_pipelines(X, y, pbar, raise_errors=raise_errors)
    368 
    369         current_batch_pipelines = []

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/evalml/automl/automl_search.py in _add_baseline_pipelines(self, X, y, pbar, raise_errors)
    469         pbar.set_description_str(desc=desc, refresh=True)
    470 
--> 471         baseline_results = self._compute_cv_scores(baseline, X, y, raise_errors=raise_errors, pbar=pbar)
    472         self._add_result(trained_pipeline=baseline,
    473                          parameters=baseline.parameters,

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/evalml/automl/automl_search.py in _compute_cv_scores(self, pipeline, X, y, raise_errors, pbar)
    486         for train, test in self.data_split.split(X, y):
    487             if isinstance(X, pd.DataFrame):
--> 488                 X_train, X_test = X.iloc[train], X.iloc[test]
    489             else:
    490                 X_train, X_test = X[train], X[test]

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/indexing.py in __getitem__(self, key)
   1766 
   1767             maybe_callable = com.apply_if_callable(key, self.obj)
-> 1768             return self._getitem_axis(maybe_callable, axis=axis)
   1769 
   1770     def _is_scalar_access(self, key: Tuple):

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/indexing.py in _getitem_axis(self, key, axis)
   2127         # a list of integers
   2128         elif is_list_like_indexer(key):
-> 2129             return self._get_list_axis(key, axis=axis)
   2130 
   2131         # a single integer

~/.pyenv/versions/feature_selection/lib/python3.8/site-packages/pandas/core/indexing.py in _get_list_axis(self, key, axis)
   2112         except IndexError:
   2113             # re-raise with different error message
-> 2114             raise IndexError("positional indexers are out-of-bounds")
   2115 
   2116     def _getitem_axis(self, key, axis: int):

IndexError: positional indexers are out-of-bounds

[dsherry]

Thanks @christopherbunn . Yes, let’s move this into the custom index epic. I think the fix for that should fix both this and #1126 .

[freddyaboulton]

Filed a comment about the long conversion time in Woodwork 79!