BUG: DataFrame to JSON failed when it with UUID

[grieve54706]

BUG: DataFrame to JSON failed when it with UUID

[asishm]

You can achieve this using the default_handler in df.to_json.

pd.DataFrame({"uuid": [uuid.uuid4()]}).to_json(default_handler=str).

Also the stdlib json library (pandas uses a vendored version of ujson iirc) also doesn't serialize uuids natively.

import json
import uuid
json.dumps({"a": uuid.uuid4()}) # raises with TypeError: Object of type UUID is not JSON serializable
json.dumps({"a": uuid.uuid4()}, default=str) # works

[Siddharth-Latthe-07]

@grieve54706 The issue you are encountering arises because pandas does not natively support serialization of uuid.UUID instances to JSON. When you attempt to serialize a DataFrame containing UUID objects using to_json(), it results in encoding errors.
to produce the expected behaviour
u can try out this:-
convert the UUID objects to their string representations before serializing the DataFrame to JSON.

import uuid
import pandas as pd

# Create a DataFrame with a UUID column
df = pd.DataFrame({"uuid": [uuid.uuid4()]})

# Convert UUID objects to strings
df['uuid'] = df['uuid'].astype(str)

# Serialize the DataFrame to JSON
json_data = df.to_json()
print(json_data)

Plz let me know if the above works
thanks

[grieve54706]

Thanks, guys. I think your suggestions all work.

I provide a tool to connect many databases and put the data into pandas for other people, so I will not know which column is UUID and databases have different data that could be dtype Object too.
I found the orjson serializes UUID to the string by default. Curious, pandas to JSON should follow RFC 4122 too?

[grieve54706]

Thank you everyone. I will use orjson to handle UUID.

import orjson
import uuid
import pandas as pd

df = pd.DataFrame({"uuid": [uuid.uuid4()]})
json_data  = orjson.loads(orjson.dumps(df.to_dict(orient="split"), option=orjson.OPT_SERIALIZE_UUID))