Build fact tables from RDBMS table

Anonymous — Fri, 28 May 2021 12:05:31 GMT

I want to let my Power BI sit on top of our SSAS in which we are designing cubes with facts and dimension.

Base is an OLTP DB where we are going to extract all the data with the SSIS and then create facts and dimensions. In this case I have a registration table which looks like this:

id	date_reg_start	process1_type	date_finished_process1	process2_type	date_finished_process2	date_reg_finished
1	25-05-2021	type1	25-05-2021	type1	28-05-2021	28-05-2021
2	26-05-2021	type23	28-05-2021	type9
3	27-05-2021	type1
4	27-05-2021
5

If someone start the registration the date_reg_start is set in the OLTP table. After that, if the user starts a type of process1 the type1 is set. If he finishes it, the date is set in the same row. the same procedure with process2. If process2 got completed, the registration_finish date will be set.

So, how will be look like the fact tables based on this oltp-registration table?

Re: Build fact tables from RDBMS table

Icey — Mon, 31 May 2021 05:42:44 GMT

Hi Anonymous ,

Based on my understanding, you can transform your data like below first:

let Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUTIy1TUw1TUyMAJxSioLUnELWiAJInFidaKVjEBCZmiajIzRdYFELYG0AhiDNIKVmGOxTQEFg5SaoCtVwFQWCwA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [id = _t, date_reg_start = _t, process1_type = _t, date_finished_process1 = _t, process2_type = _t, date_finished_process2 = _t, date_reg_finished = _t]), #"Unpivoted Columns" = Table.UnpivotOtherColumns(Source, {"id", "date_reg_start", "date_finished_process1", "date_finished_process2", "date_reg_finished"}, "Attribute", "Value"), #"Unpivoted Columns1" = Table.UnpivotOtherColumns(#"Unpivoted Columns", {"id", "date_reg_start", "date_reg_finished", "Attribute", "Value"}, "Attribute.1", "Value.1"), #"Replaced Value" = Table.ReplaceValue(#"Unpivoted Columns1","_type","",Replacer.ReplaceText,{"Attribute"}), #"Replaced Value1" = Table.ReplaceValue(#"Replaced Value","date_finished_","",Replacer.ReplaceText,{"Attribute.1"}), #"Added Conditional Column" = Table.AddColumn(#"Replaced Value1", "Custom", each if [Attribute] = [Attribute.1] then 1 else null), #"Filtered Rows" = Table.SelectRows(#"Added Conditional Column", each ([Custom] = 1)), #"Removed Columns" = Table.RemoveColumns(#"Filtered Rows",{"Custom", "Attribute.1"}), #"Renamed Columns" = Table.RenameColumns(#"Removed Columns",{{"Attribute", "process"}, {"Value", "type"}, {"Value.1", "date_finished"}}), #"Changed Type with Locale" = Table.TransformColumnTypes(#"Renamed Columns", {{"date_reg_start", type date}, {"date_reg_finished", type date}, {"date_finished", type date}}, "en-GB"), #"Changed Type" = Table.TransformColumnTypes(#"Changed Type with Locale",{{"id", Int64.Type}, {"date_reg_start", type date}, {"date_reg_finished", type date}, {"process", type text}, {"type", type text}, {"date_finished", type date}}), #"Reordered Columns" = Table.ReorderColumns(#"Changed Type",{"id", "date_reg_start", "process", "type", "date_finished", "date_reg_finished"}) in #"Reordered Columns"

Then, you can create a Date dimension table and process, process type dimension table according to your requirements.

If this is not what you want, please share me more details.

Best Regards,

Icey

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

topic Build fact tables from RDBMS table in Report Server

Build fact tables from RDBMS table

Re: Build fact tables from RDBMS table