zhao, zhenhui

> Hi Zhenhui, looks like a custom format "dwrf" to Substrait is not supported. If running with "spark.read.format("dwrf").load", below error occurs. > > ``` > Reason: ReaderFactory is not registered...

> @zhejiangxiaomai These changes are hard to follow. Would it be possible to split this PR in to two: add support for aggregations + add support for q6. It would...

Hi, @mbasmanova In my opinion, all comments were fixed. Do you review again? and merge it.

> @zhejiangxiaomai The PR title says "add support for aggregations", which suggests that without the changes in the PR plans with aggregations were not supported. Is this the case? If...

> @zhejiangxiaomai Thank you for clarify the PR title. It reads "support avg function based on row_construct" now. Would you also update the PR description to explain what does that...

Hi,@mbasmanova I think all comments were fixed, could you please review it again when you have time? Actually, I have to clarify some points about this PR. 1. A round-trip...

> @zhejiangxiaomai Looks good overall % some comments below. > > > we found transfer "row_construct" to substrait plan is not supported now > > I'm curious what's the problem...

> @zhejiangxiaomai Looks good overall % a couple of questions. > > I find it difficult to understand the PR title and description and wonder if there is a way...

> @zhejiangxiaomai I spent some time running and debugging Substrait conversion code. > > First, I ran into this error: > > ``` > error: 'args' is deprecated [-Werror,-Wdeprecated-declarations] >...

> @zhejiangxiaomai > > > I will make a PR to fix "args error" > > That will be awesome. Thank you! > > > Then we can merge #2264...