OpenMLDB
OpenMLDB copied to clipboard
Real-time feature calculation for full table
Describe the feature you'd like
+---+-----+---+-------+-------------------+---------+----------+--------+--------+-----+
| id| uid|mid| cardno| trans_time|trans_amt|trans_type|province| city|label|
+---+-----+---+-------+-------------------+---------+----------+--------+--------+-----+
| 1|UserA| M1|1111111|2022-02-02 10:00:00| 10.0| CASH| SH|ShangHai| true|
| 3|UserA| M2|1111111|2022-02-03 11:00:00| 20.0| POS| SH|ShangHai| true|
| 4|UserA| M3|1111111|2022-02-07 11:00:00| 15.0| POS| SH|ShangHai| true|
| 6|UserA| M1|1111111|2022-02-07 13:00:00| 20.0| CASH| BJ| BeiJing| true|
| 8|UserA| M2|1111111|2022-02-08 11:00:00| 50.0| CASH| BJ| BeiJing| true|
| 9|UserA| M3|1111111|2022-02-08 12:00:00| 100.0| POS| BJ| BeiJing| true|
| 2|UserB| M3|2222222|2022-02-01 11:30:00| 10.0| POS| SH|ShangHai| true|
| 5|UserB| M1|2222222|2022-02-07 10:00:00| 20.0| POS| SH|ShangHai| true|
| 7|UserB| M2|2222222|2022-02-07 12:00:00| 30.0| POS| SH|ShangHai| true|
| 10|UserB| M5|2222222|2022-02-08 11:00:00| 10.0| POS| SH|ShangHai| true|
+---+-----+---+-------+-------------------+---------+----------+--------+--------+-----+
比如:查看city最近10分钟出现最多的城市 这种deployment是全表统计的,实时特征计算应当无需任何input,而当前的deployment的partition字段和请求的input(类似{"input":[{{rowdata}}],"need_schema":false})字段是必须 请问是否考虑这种实时特征计算的支持呢
Additional context
Add any other context or screenshots about the feature request here.