pitrou commented on PR #14320:
URL: https://github.com/apache/arrow/pull/14320#issuecomment-1372473478
One other possibility would be to first create the function and then
register kernels one by one:
```python
udf = pc.register_scalar_function(
name="y=x*2",
arg_names=["x"],
doc = {"summary": "multiply by two function",
"description": "test multiply function"}
)
for in_type in [pa.int64(), pa.float64()]:
udf.add_kernel([in_type], in_type, impl)
```
It could also open the way to a decorator-like syntax:
```python
udf = pc.register_scalar_function(
name="y=x*2",
arg_names=["x"],
doc = {"summary": "multiply by two function",
"description": "test multiply function"}
)
@udf.kernel([pa.float64()], pa.float64())
@udf.kernel([pa.int64()], pa.int64())
def impl(ctx, x):
# ...
```
@jorisvandenbossche Opinions on which kind of API would be desirable here?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
