how to use "select concat(col)" in siddhi aggregation query?

Question

How to get all data of one column in the aggregation siddhi query. For example, i have the data as:

column1 column2 column_uuid
1       a       uuid1
2       a       uuid1
3       a       uuid3
4       b       uuid4

I want to use the siddhi query as:

define stream Input (column1 int, column2 string, column_uuid string);

define stream Output (column2 string, amount long, uuid string);

@info(name='query')
from Input#window.time(30 sec)
select column2, count() as amount, concat(column_uuid) as uuid
group by column2
having amount > 2
insert into Output;

and i want get result as:

Event{timestamp=xxx, data=[a, 3, "uuid1,uuid2,uuid3"]}

Answer 1

In Siddhi, there are two types of functions. Regular "function"s, and "aggregate-function"s. Since str:concat() is not an aggregate-function, you can't use that to get your desired outcome.

Therefore, you may have to write your own custom:concat() aggregate function to get that expected result. Please refer to following samples on writing a custom aggregate functions (samples). To get the above output, you can simply keep a global string variable within your custom attribute aggregator and append to that in processAdd(Object data) method (similar to this).