how to add udf to chdb in the easy way?

[l1t1]

I saw https://ucasfl.github.io/2020/12/25/How-to-Implement-User-defined-functions-in-ClickHouse/

it needs add new files of the udf and modify the source code of clickhouse,

https://datafusion.apache.org/library-user-guide/functions/adding-udfs.html seems easier, it only need add new files.

what is the easy way for chdb?

[auxten]

Here is an example:

#!/usr/bin/env python

from chdb.udf import chdb_udf
from chdb import query

@chdb_udf()
def sum_udf(lhs, rhs):
    return int(lhs) + int(rhs)

print(query("select sum_udf(12,22)"))

[l1t1]

thanks, but python udf is a bit slow.

>>> import time
>>> t=time.time();print(query("select sum(cast (sum_udf(i,j) as int)) from(select number i,number*3 j from numbers(1000000))"));print(time.time()-t)
1999998000000

2.1727423667907715
>>> t=time.time();print(query("select sum(cast (i+j as int)) from(select number i,number*3 j from numbers(1000000))"));print(time.time()-t)
1999998000000

0.03494834899902344