Performance Issues with Group By Clause

Pragati11 If that is the case just use CONCATENATEX DAX function instead of doing group by in PQ

Pragati11 I would still lean towards keeping this in Power Query. You can do the concatenation (TEXT.COMBINE) in the Group By function. I just did this yesterday for a client.

But - performance wasn't great. Added 30min to the transformation. So I moved this to a Dataflow, and let it to the heavy lifting. Still takes about 30min, but it is offloaded from Power BI and doesn't impact your development. If you have other heavy transformations, if you move them to dataflows as well, they can run in parallel, and when you load Power BI data, it is just a straight load of materalized tables. 

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WSlTSUQpSitWJVkoCsgLALJCYF1zMAy6GkPWHsxzh6lzhrFA4KwSuLhwu5oxmXiqQ5QSX9QazkuEmpwBZfnB1wXBZT7gs0H2xAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Group = _t, Code = _t]),
    #"Grouped Rows" = 
        Table.Group(
            Source, 
            {"Group"}, 
            {
                {
                    "All Rows", 
                    each Text.Combine(_[Code], ", "), 
                    type table [Group=nullable text, Code=nullable text]
                }
            }
        )
in
    #"Grouped Rows"

That code converts this:

to this in one step.

Given the amount of data you are talking about, pushing more back to the source can make the model perform better since the vertipaq engine can optimize the compression if the data needs to be in a column. If it just needs to be in a measure, then yeah, CONCATENATEX() will work - but you will have a lot more records as you didn't group before loading.

So.... it depends. 😁

How to use M code provided in a blank query:
1) In Power Query, select New Source, then Blank Query
2) On the Home ribbon, select "Advanced Editor" button
3) Remove everything you see, then paste the M code I've given you in that box.
4) Press Done
5) See this article if you need help using this M code in your model.

Pragati11 What type is your original data source?
Can you already do the grouping on the team table before the join?

What are the cases in which you cannot avoid the computed entites?

Pragati11 Thanks for tagging. Just to clarify, you raw data contains two tables:

1 - Project Table with unique projects

2 - Team table with multiple team members for each projects 

and above two tables can be joined on ProjectId which will be 1 to many relationship, correct?

And you want the output to show the Team members of each project separated by a comma, correct?

Please confirm when you get a chance.