Hi, Tucker
IIUC, this way is unsupported. You may try other way to work around it.
--
Best！
Xuyang
At 2024-05-17 00:22:57, "Tucker Harvey" <thar...@pinterest.com.INVALID> wrote:
>Hi Xuyang,
>
>Thanks for your response. I can provide some better context, we are migrating
>a use case to FlinkSQL and they want to do some query that has a UDF like
>
>SELECT user_id,
> ARRAY_MAP(transactions, x -> x * 1.1) AS taxed_transactions
>FROM users;
>
>Based on my understanding of the API for UDFs in Flink, this will not be
>possible with the current framework?
>
>Best,
>Tucker
>
>
>> On May 15, 2024, at 7:42 PM, Xuyang <xyzhong...@163.com> wrote:
>>
>> Hi, Tucker.
>>
>> Could you provide some examples of how you use lambda expressions?
>>
>> Currently, if you try to use a lambda expression as a member variable of a
>> UDF class, it will fail because UDF
>>
>> classes are required to be serializable, and lambda expressions are not
>> serializable.
>>
>>
>>
>>
>> However, there are workarounds, such as marking the lambda member variable
>> function with the transient modifier
>>
>> to prevent it from being serialized, and initializing the function in the
>> overridden open method.
>>
>>
>>
>>
>> I've provided an example below:
>>
>>
>>
>>
>> ```
>>
>> public class JavaFunc22 extends ScalarFunction {
>>
>>
>>
>>
>> transient Function<String, String> f;
>>
>>
>>
>>
>> @Override
>>
>> public void open(FunctionContext context) throws Exception {
>>
>> f = str -> str + ":test";
>>
>> }
>>
>>
>>
>>
>> public String eval(String s) {
>>
>> return f.apply(s);
>>
>> }
>>
>> }
>>
>>
>>
>>
>> ```
>>
>>
>>
>>
>> --
>>
>> Best！
>> Xuyang
>>
>>
>>
>>
>>
>> 在 2024-05-15 23:49:50，"Tucker Harvey" <thar...@pinterest.com.INVALID> 写道：
>>> Hi Flink Community,
>>>
>>> I’m writing to confirm whether lambda expressions are supported with User
>>> Defined Functions (UDFs) in FlinkSQL and the Table API. My current
>>> understanding is that they are not supported.
>>> Can anyone verify this, or let me know if there have been any recent
>>> changes regarding this?
>>> Thanks for your help.
>>>
>>> Best regards,
>>> Tucker
>
