Copy keys to a new dictionary (Python)

Question

I’m reading a csv file, using DictReader(). The function returns a dictionary, where the header items are the keys and the cells are the values. Pretty cool.

But I’m trying to account for rows where the data may not be what I expect it to be. In that case (I’m catching a ValueError exception), I would like the rows that are ‘suspect’ to go into a separate dictionary, for manual processing.

My question is this: since my first dictionary (the object returned by DictReader) has all of its keys set up properly, how do I copy just the keys into my second dictionary, the one which I want to be just a dictionary of suspect rows, to be manually processed?

I’ve been toying around with dict.fromkeys() and such for a while now and I’m just not getting anywhere. Halp!

EDIT: Pasting some of my erroneous code. Going to go hide in shame of my code. Don’t judge me! 😉

unsure_rows = dict.fromkeys(dict(csv_reader).keys(), [])
for row in csv_reader:
#   if row['Start Time'] != 'None':
    try:
        if before_date > strptime(row['Start Time'], '%Y-%m-%d %H:%M:%S') > after_date:
            continue
    except ValueError:
        unsure_rows += row

ValueError: dictionary update sequence element #0 has length 13; 2 is required

Asked By: Harv

||

Source

Answer 1

You wouldn’t want to copy “just the keys” into another dictionary, but if you have “just the keys” you will have a set.

To get the keys from dict d, you need only say d.keys().

This returns a list (with keys in arbitrary order), which you can keep as a list or copy into a set with

set(d.keys())

Example:

>>> d = {'one': 1, 'three': 3, 'two': 2, 'four': 4} 
>>> set(d.keys())
set(['four', 'one', 'three', 'two'])

EDIT:

Now I see that you intend to capture suspect key-value pairs as you catch exceptions. In this case, just start with an empty dictionary

suspect = {}

And inside your code, which I would imagine is some kind of loop, add the suspect key value pairs like so:

while something():
    k, v = generate_pair()
    try:
        analyze_something_with_k_and_v_that_might_throw_an_exception
        add_it_to_regular_dict
    except:
        suspect[k] = v

Answered By: Ray Toal

Answer 2

Have you tried:

newd = dict.fromkeys(origdict)

? If that doesn’t work for you then please add more details about the error you are getting.

Answered By: Keith

Answer 3

You are close. Try:

dict.fromkeys(my_csv_dict.keys(),[])

This will initialize a dictionary with the same keys that you parsed from your CSV file, and each one will map to an empty list (to which, I assume, you will append your suspect row values).

Try this. (There are several subtler changes here that are all necessary, like how you can’t initialize unsure_rows before you start reading the CSV.)

unsure_rows = None
for row in csv_reader:
#   if row['Start Time'] != 'None':
    try:
        if before_date > strptime(row['Start Time'], '%Y-%m-%d %H:%M:%S') > after_date:
            continue
    except ValueError:
        if not unsure_rows:
            # Initialize the unsure rows dictionary
            unsure_rows = dict.fromkeys(csv_reader.fieldnames,[])
        for key in unsure_rows:
            unsure_rows[key].append(row[key])

Answered By: cheeken

Answer 4

The solution from cheeken isn’t working anymore as dict.fromkeys is creating the dictionary with all values pointing to the same list you should instead use {k: [] for k in my_csv_dict.keys()}

Answered By: ajreckof

Copy keys to a new dictionary (Python)

Question:

Answers: