How to Convert a String to a Dictionary in Python

To convert a string to a dictionary in Python, use ast.literal_eval() when the text looks like a Python dict (single quotes, True, None), and json.loads() when it’s JSON. Both return a real dict:

import ast, json
ast.literal_eval("{'state': 'TX', 'total': 49.99}")   # {'state': 'TX', 'total': 49.99}
json.loads('{"state": "TX", "total": 49.99}')         # same dict

I tested these conversions on Python 3.12.5 from Command Prompt, feeding them made-up US customer orders.

The right tool depends on what the string looks like, so we’ll go format by format and finish with a safe helper.

Convert a Python dict string with ast.literal_eval()

A string like "{'customer': 'Emily Carter'}" usually comes from str(my_dict) or a log file. ast.literal_eval() reads it back, nested lists and booleans included:

import ast

text = "{'customer': 'Emily Carter', 'state': 'TX', 'balance': 1250.5, 'active': True, 'tags': ['pro', 'annual']}"

order = ast.literal_eval(text)
print(type(order))
print(order["customer"], order["tags"][0])
print(f"Balance: ${order['balance']:,.2f}")

Output:

<class 'dict'>
Emily Carter pro
Balance: $1,250.50

You can trust it because it only accepts literals, never function calls. Going the other way is covered in converting a dictionary to a string in Python.

Convert a JSON string to a dictionary with json.loads()

API responses use JSON: double quotes, true and null. json.loads() maps those to True and None, but a single-quoted Python string makes it fail:

import ast
import json

api_text = '{"customer": "James Miller", "total": 49.99, "paid": true, "coupon": null}'
data = json.loads(api_text)
print(data)

repr_text = "{'customer': 'James Miller', 'paid': True}"
try:
    json.loads(repr_text)
except json.JSONDecodeError as err:
    print("json.loads failed:", err)

print(ast.literal_eval(repr_text))

Output:

{'customer': 'James Miller', 'total': 49.99, 'paid': True, 'coupon': None}
json.loads failed: Expecting property name enclosed in double quotes: line 1 column 2 (char 1)
{'customer': 'James Miller', 'paid': True}

Don’t swap the quotes with replace(); you’ll break names like O’Brien. Nested JSON, dates and JSON files are in my guide to converting a JSON string to a dictionary.

How do I convert a key=value string to a dictionary?

Text such as name=Olivia Brown; state=CO isn’t a literal, so split it yourself: on the pair separator, then on the first =, stripping spaces:

line = "name=Olivia Brown; city=Denver; state=CO; plan=pro; total=79.5"

record = {}
for pair in line.split(";"):
    key, value = pair.split("=", 1)      # maxsplit=1 keeps any '=' inside the value
    record[key.strip()] = value.strip()
print(record)

record["total"] = float(record["total"])
print(f"{record['name']} owes ${record['total']:.2f}")

# one-liner for simple "key:value,key:value" text
codes = dict(item.split(":", 1) for item in "TX:Texas,CA:California,NY:New York".split(","))
print(codes)

Output:

{'name': 'Olivia Brown', 'city': 'Denver', 'state': 'CO', 'plan': 'pro', 'total': '79.5'}
Olivia Brown owes $79.50
{'TX': 'Texas', 'CA': 'California', 'NY': 'New York'}

You get strings from split(), so convert numbers yourself. More splitting options are in splitting a string by comma in Python, and extra spaces are handled in removing spaces from a string.

Convert a URL query string to a dictionary

For state=TX&plan=pro text from a URL, the standard library parses and decodes it. parse_qsl() gives pairs for dict(); parse_qs() keeps repeated keys:

from urllib.parse import parse_qs, parse_qsl, urlsplit

url = "https://shop.example.com/checkout?state=TX&plan=pro&total=49.99&tag=new&tag=sale"
query = urlsplit(url).query

print(dict(parse_qsl(query)))   # one value per key; the last tag wins
print(parse_qs(query))          # every value kept in a list

Output:

{'state': 'TX', 'plan': 'pro', 'total': '49.99', 'tag': 'sale'}
{'state': ['TX'], 'plan': ['pro'], 'total': ['49.99'], 'tag': ['new', 'sale']}

Use parse_qs() when a key can repeat, like the two tag values here. Then a dictionary comprehension can convert the values.

Safely convert any string to a dict, and why not eval()

eval() works too, but it runs any code in the string, so I never use it on text I didn’t write. This helper tries JSON, then literal_eval(), and rejects the rest:

import ast
import json

def to_dict(text):
    """Turn JSON or a Python dict literal into a dict, or raise ValueError."""
    try:
        result = json.loads(text)
    except json.JSONDecodeError:
        try:
            result = ast.literal_eval(text)
        except (ValueError, SyntaxError) as err:
            raise ValueError(f"not a dictionary string: {err}") from None
    if not isinstance(result, dict):
        raise ValueError(f"expected a dict, got {type(result).__name__}")
    return result

samples = ['{"plan": "pro", "seats": 5}',
           "{'plan': 'team', 'seats': 12}",
           "__import__('os').getcwd()",
           "[1, 2, 3]"]
for s in samples:
    try:
        print("OK  ", to_dict(s))
    except ValueError as err:
        print("FAIL", err)

Output:

OK   {'plan': 'pro', 'seats': 5}
OK   {'plan': 'team', 'seats': 12}
FAIL not a dictionary string: malformed node or string on line 1: <ast.Call object at 0x000002216ACA4DD0>
FAIL expected a dict, got list

You need the isinstance() check because literal_eval() also returns lists. If you later need key-value pairs, see converting a dictionary to a list of tuples.

String to dictionary methods compared

String looks like	Use	Watch out for
{'a': 1, 'ok': True}	ast.literal_eval()	Returns any literal, so check for dict
{"a": 1, "ok": true}	json.loads()	Single quotes raise JSONDecodeError
a=1; b=2	split() + dict()	Values stay strings
a=1&b=2	parse_qsl() / parse_qs()	Repeated keys

You’ll probably want these next:

• Convert a string to JSON in Python
• Convert a string to an integer in Python
• Handle a KeyError with try/except in a dictionary
• Convert a dictionary to JSON in Python

Frequently asked questions