For example.
i. Table named test which contains two column named field_a and field_b.
ii. Than I load csv file. The script don’t contains column field_a.
iii. But when I query the table result. I find the column field_a still exist.
iv. If I remove format of the table firstly and than load csv file again. Finally the column will be deleted.
Hi, Community. Above the phenomenon, is it an issue.
@mccann.ren
Since Dremio learnt about field_a in the beginning, it remembers it, you can try to do a metadata refresh and if that does not help try a FORGET metadata
@balaji.ramaswamy
Kindly thanks for your answer.
It is exactly that refresh metadata is useless.
Maybe I could try to forget metadata.
@mccann.ren
You said the below
- Table named test which contains two column named field_a and field_b.
- Then I load csv file. The script don’t contains column field_a.
Since the table has field_a Dremio has to learn it, if the incoming file does not have field_a then that’s schema learning but Dremio cannot forget field_a as it is still part of the dataset