How to use dtype inside a for loop while traversing a dataframe column-wise ? So that I can traverse-CodePudding

For example, I have a dataframe 'df' like this:

Name	color	id	weight
john	blue	67	70
clara	yellow	-	67
diana	red	89	56

Here the numeric columns like "id" and "weight" should have all numeric values, unlike the second value of "id" which is a '-'.

If I do df.dtypes, it returns:

| name | object

**How can I traverse through the dataframe column-wise, then check if the type of column is an object, then if it an object, then check if it is becoming an object because of the typo '-' like id- if yes then raise a flag **

CodePudding user response：

Zip up the column name and the dtypes to make a tuple:

for col_name, col_type in zip(df.columns, df.dtypes):
    if col_type == "object":
        # do whatever here
        pass

CodePudding user response：

import numpy as np

object_cols = df.select_dtypes("object").columns

result = []

for col in object_cols:
  try:
    df[col].replace("-",0).astype("int64") 
    result.append(col)
  except:
     pass

and result would contain all columns that could be of type int if there is no "-"