qingzma / dbestclient Goto Github PK

Hi Qingzhi,

I am fascinated by your project and want to check how it works.
Here is what I have done.

Could you please, explain why the result is not like in the dbestclient?
I want to see 578.38

Regards
Fedor

Hi.
I used DBEst for approximate query processing. I found that a model only supports two attributes. Does that mean I should create more models if there are more than two attributes in the query ?

Hi qingzhi,
The paper mentions that the queries are randomly generated. Could you offer me the TPC-DS workload, CCPP workload and Beijing PM2.5 workload? How do you randomly generate the queries and what are the queries like? I am really grateful if you could tell me the details. My email: [email protected].
By the way, you use relative error as metric, what is the definition of relative error? How to calculate ?
Thanks.

Welcome to DBEst: a model-based AQP engine! Type exit to exit!
dbestclient> create table mdl(pm25 real, PRES real) from pm25.csv method uniform size 1000
dbestclient> select count(pm25 real) from mdl where 1000<=PRES<=1020;
Start creating model mdl
Reading data file...
Traceback (most recent call last):
File "D:\MyLoad\python3.8.2\lib\cmd.py", line 214, in onecmd
func = getattr(self, 'do_' + cmd)
AttributeError: 'DBEstPrompt' object has no attribute 'do_select'

During handling of the above exception, another exception occurred:
这个是为什么啊？