Detecting the type of data we have

Question

I am importing a csv file into python. As expected, it creates lists with every value as a string, however I want to avoid this. Is there a way for python to detect that a value is actually an int even though it looks like this '24'?

Thank you.

What library are you using? what does your data look like? You have to give more details — João Almeida, Nov 05 '16 at 14:49
what if some column contains sometimes a string and sometimes an int in a string? you have to be more specific: show us the code. — Jean-François Fabre, Nov 05 '16 at 14:50
See [this](http://stackoverflow.com/questions/15197673/using-pythons-eval-vs-ast-literal-eval) answer. — Rockybilly, Nov 05 '16 at 14:54

saurabh baid · Accepted Answer · 2016-11-05T15:32:57.980

1

You may want to write a function something like this to do the job. You can expand it to cover other data type.

def return_str_type(str):
    possible_type = [int, float]
    for dtype in possible_type:
        try:
            str = dtype(str)
            break   
        except:
            pass
    return type(str).__name__

print(return_str_type('4'))
print(return_str_type('4.3'))
print(return_str_type('s4'))

This will give OP

int
float
str

Though you will have to be careful with the order. e.g. int check should always be before float.

edited Nov 05 '16 at 15:32

answered Nov 05 '16 at 15:21

saurabh baid

1,819
1
14
26

you don't need eval for this, make a list with the classes and do `str=dtype(str)` – Copperfield Nov 05 '16 at 15:29
updated as per you suggestion @Copperfield – saurabh baid Nov 05 '16 at 15:33
You beat me to it. I had the same except the function excepted a complete row and returned a list of converted items. Used with ```data = map(f, csv.reader)``` – wwii Nov 05 '16 at 15:36
@saurabhbaid I think you're onto something! I managed to get a whole list of str and int. Thanks so much! – plshelp Nov 05 '16 at 15:45
1

As your example shows, there's a built-in type name `str`, so it's best to avoid defining your own variable, arguments, etc, the same. – martineau Nov 05 '16 at 16:59

Rockybilly · Answer 2 · 2016-11-05T15:04:06.263

0

ast.literal_eval is what you seek.

print ast.literal_eval('[1, 2, 3]') == [1, 2, 3]
print ast.literal_eval('23') == 23

Python also has an eval command, however it is not safe because non-datatype commands can also be processed for harmful purposes through this function.

edited Nov 05 '16 at 15:04

answered Nov 05 '16 at 14:58

Rockybilly

2,938
1
13
38

Thank you, I've given it a go and it says Value error: Malformed string but I'll keep trying. – plshelp Nov 05 '16 at 15:05

score -1 · Answer 3 · answered Nov 05 '16 at 14:49

-1

A number of various modules supports this, pandas among them. read_csv can infer the data types.

answered Nov 05 '16 at 14:49

Lukasz Tracewski

10,794
3
34
53

Detecting the type of data we have

3 Answers3