DataFrame.eval error fixed #59145
DataFrame.eval error fixed #59145
Conversation
| @@ -218,3 +217,12 @@ def reconstruct_object(typ, obj, axes, dtype): | |||
| ret_value = np.array([ret_value]).astype(res_t) | |||
|
|
|||
| return ret_value | |||
|
|
|||
| # Test with the DataFrame and multiline expression | |||
There was a problem hiding this comment.
Can you move the test to a proper test file? I think pandas/tests/test_expressions.py is probably the best place for it.
| @@ -144,8 +145,6 @@ def _align_core(terms): | |||
| obj = ti.reindex(reindexer, axis=axis) | |||
| terms[i].update(obj) | |||
|
|
|||
| terms[i].update(terms[i].value.values) | |||
There was a problem hiding this comment.
Removing this line does fix the bug, nice job! But I'm curious about potential side effects: this line of code converts terms[i].value from pd.Series to np.ndarray, and removing it might affect the behavior of terms in whatever context it is used. Are the test failures related?
There was a problem hiding this comment.
Sure @chaoyihu I have added the test in the respective file
But here what you want me to do further in this terms[i].value ??
There was a problem hiding this comment.
I think we should wait and see if it passes the checks after this revision.
There was a problem hiding this comment.
Thanks for the quick response! I viewed your changes and left new comments.
The check failures seem to have been caused by the import line you added to the beginning of align.py - from __future__ imports must occur at the beginning of the file.
You probably should also rebase your branch on main, since your branch is significantly behind.
For now, I do not think you need to do anything further with terms[i]. Can you revise the code, do a rebase, then push again and wait to see if it passes the checks?
If it does, someone with write access will likely come to inspect and merge your code.
|
I have moved the test in the respective file |
| @@ -2,6 +2,7 @@ | |||
| Core eval alignment algorithms. | |||
| """ | |||
|
|
|||
| import pandas as pd | |||
| @@ -218,3 +218,12 @@ def reconstruct_object(typ, obj, axes, dtype): | |||
| ret_value = np.array([ret_value]).astype(res_t) | |||
|
|
|||
| return ret_value | |||
|
|
|||
| # Test with the DataFrame and multiline expression | |||
There was a problem hiding this comment.
Please don't forget to remove this block of code from align.py, since you've added it to the test file.
| @@ -2,6 +2,7 @@ | |||
| import re | |||
|
|
|||
| import numpy as np | |||
| import pandas as pd | |||
| # Define expected values manually or calculate based on expectation | ||
| expected_first_ret = df['first'] / df['pre'].fillna(df['first']) - 1.0 | ||
| expected_last_ret = df['last'] / df['pre'].fillna(df['first']) - 1.0 | ||
| expected = pd.DataFrame({"first_ret": expected_first_ret, "last_ret": expected_last_ret}) |
There was a problem hiding this comment.
Please use DataFrame instead of pd.DataFrame.
|
|
||
| # New test added here | ||
| def test_multiline_expression(self): | ||
| df = pd.DataFrame({"first": [9.76, 9.76, 9.76], "last": [9.76, 9.76, 9.76], "pre": [9.75, 9.76, 9.76]}) |
There was a problem hiding this comment.
Please use DataFrame instead of pd.DataFrame.
|
I am closing this PR |
DataFrame.evalfails with TypeError with multiline expr but works whenevalline by line #59062 (Replace xxxx with the GitHub issue number)doc/source/whatsnew/vX.X.X.rstfile if fixing a bug or adding a new feature.