- Operating system: Ubuntu 20.04
- SonarLint plugin version: 3.22.0
- Programming language you’re coding in: Python
- Is connected mode used: Connected to SonarCloud
Hello,
I’ve encountered a potential false positive with SonarLint when working with pandas DataFrames in Python. Here’s a simplified working example of the code:
import pandas as pd
# Sample data
bar_list = ['A', 'B', 'C']
foo_df = pd.DataFrame({
'FOO': ['A', 'B', 'C'],
'BAR': ['B', 'C', 'A'],
'VALUE1': [1, 2, 3],
'VALUE2': [4, 5, 6]
})
foo_key = 'FOO'
bar_key = 'BAR'
def main():
foo_matrix = pd.DataFrame(index=bar_list, columns=bar_list)
bar_matrix = pd.DataFrame(index=bar_list, columns=bar_list)
for _, row in foo_df.iterrows():
foo_entry = row.VALUE1
bar_entry = row.VALUE2
foo_value = row.VALUE3
bar_value = row.VALUE4
foo_matrix.loc[foo_entry, bar_entry] = foo_value
bar_matrix.loc[foo_entry, bar_entry] = bar_value
Screenshot of the problem:
The following lines are flagged
foo_matrix.loc[foo_entry, bar_entry] = foo_value
bar_matrix.loc[foo_entry, bar_entry] = bar_value
with a warning that the value set in the first line might be overwritten by the second line. However, these lines are writing to two different DataFrames (foo_matrix and bar_matrix), so they cannot overwrite each other’s values.
Is that intended? A workaround to avoid the warning is the following, but it doesn’t feels right.
for _, row in foo_df.iterrows():
foo_matrix.loc[row.VALUE1, row.VALUE2] = row.VALUE3
bar_matrix.loc[row.VALUE1, row.VALUE2] = row.VALUE4