1. import pandas as pd
  3. def normalize_metadata(df, columns_to_normalize):
  4. """
  5. Normalizes metadata columns in a pandas DataFrame.
  7. Args:
  8. df (pd.DataFrame): The input DataFrame containing metadata.
  9. columns_to_normalize (list): A list of column names to normalize.
  11. Returns:
  12. pd.DataFrame: The DataFrame with normalized metadata. Returns original df if columns_to_normalize is empty.
  13. """
  15. if not columns_to_normalize:
  16. return df # Return original DataFrame if no columns to normalize
  18. for col in columns_to_normalize:
  19. if col in df.columns:
  20. # Convert column to numeric if possible
  21. try:
  22. df[col] = pd.to_numeric(df[col], errors='coerce')
  23. except ValueError:
  24. pass #if conversion fails, keep as is
  26. # Normalize the column (example: min-max scaling)
  27. min_val = df[col].min()
  28. max_val = df[col].max()
  30. if max_val - min_val > 0: # Avoid division by zero
  31. df[col] = (df[col] - min_val) / (max_val - min_val)
  32. else:
  33. df[col] = 0 # If min and max are equal, set to zero
  34. return df
  36. if __name__ == '__main__':
  37. # Example Usage
  38. data = {'col1': [10, 20, 30, 40, 50],
  39. 'col2': ['a', 'b', 'c', 'd', 'e'],
  40. 'col3': [1, 2, 3, 4, 5]}
  41. df = pd.DataFrame(data)
  43. columns_to_normalize = ['col1', 'col3'] # specify columns for normalization
  45. normalized_df = normalize_metadata(df.copy(), columns_to_normalize) # create a copy to avoid modifying original
  47. print(normalized_df)

Add your comment