Feature Type

• Adding new functionality to pandas

• Changing existing functionality in pandas

• Removing existing functionality in pandas

Problem Description

Hey Pandas team,

Currently, the eval method treats special characters in column names different than the query method.
The query method allows for the use of backticks to enclose column names with special characters in contrast to the eval method. This inconsistency can be confusing for users.

The proposal is to enhance the parsing capabilities of the eval method to recognize and handle backticks in a manner similar to the query method. This modification would provide users with more flexibility when working with DataFrame expressions, especially in cases where column names contain special characters.

Here is an example:

import pandas as pd

df = pd.DataFrame({'my.col': [1, 2, 3]})
qry = df.query('`my.col` > 2')
evl = df.eval('`my.col` = 0')
print('Original:')
print(df)
print('\nquery: `my.col` > 2')
print(qry)
print('\neval: `my.col` = 0')
print(evl)

# Original:
#    my.col
# 0       1
# 1       2
# 2       3

# query: `my.col` > 2
#    my.col
# 2       3

# eval: `my.col` = 0
#    my.col  BACKTICK_QUOTED_STRING_my_DOT_col
# 0       1                                  0
# 1       2                                  0
# 2       3                                  0

# Current behavior in eval
# Creates a column with modified name 'BACKTICK_QUOTED_STRING_my_DOT_col'

Feature Description

# Proposed behavior in eval
# Directly modifies the 'my.col' column
# df.eval('`my.col` = 0')
#    my.col
# 0       0
# 1       0
# 2       0

Alternative Solutions

Can eval parsing be adapted to the way query parses?

Additional Context

Python 3.11.5
Pandas 2.1.1