DAX: Merge two tables selectively (based on a value in a row) - merge

Question

I have two tables:
Table with Division and Entity within divisions
Table with projects identified per entity or for all entities in a division.
**Table 1: Organization**
id Division Entity
- - - - - - - - - - - - -
0 Europe France
1 Europe Germany
2 Europe Italy
3 Europe Spain
4 China North
5 China East
6 China West
7 China South
**Table 2: Project**
id Division Org. Project
- - - - - - - - - - - - - - - - - - - - - - - - - -
1 Europe France Project 1
2 Europe Germany Project 2
3 Europe Germany Project 3
4 Europe All entities Project 4
5 Europe All entities Project 5
6 China East Project 6
7 China All entities Project 7
Some projects are specific to [Org.]'s, while others are relevant for all [Org.]'s in the [Division]. In these cases the value "All entities" is used.
I want to merge both tables based on [Division], but just for Plan[Org.] = "All entities", while the rows with other values (e.g. "France", "Germany") should just be repeated.
The resulting table should look like this:
**Table 3: Merge**
id Division Org. Entity Project
- - - - - - - - - - - - - - - - - - - - - - - - - -
1 Europe France France Project 1
2 Europe Germany Germany Project 2
3 Europe Germany Germany Project 3
4 Europe All entities France Project 4
4 Europe All entities Germany Project 4
4 Europe All entities Spain Project 4
4 Europe All entities Italy Project 4
5 Europe All entities France Project 5
5 Europe All entities Germany Project 5
5 Europe All entities Spain Project 5
5 Europe All entities Italy Project 5
6 China East East Project 6
7 China All entities North Project 7
7 China All entities East Project 7
7 China All entities South Project 7
7 China All entities West Project 7
A physical relationship between the tables doesn't work because of the missing "All entities" value in the Organization table, but it can be added when with an additional [Entity id] column in the Project table, if needed.
I prefer DAX. Solutions in PowerQuery are less preferred as I am not so fluent in M, but still welcome if
available.
Thank you!

Answer 1

Follow these below steps in Power Query-
Step-1: Create a custom column* in table Organization as below-
Here is the output-
Step-2: Create 2 Duplicate table from table Project and name them as - Project_all_emtities and Project_others
Step-3: Go to Advanced Editor for table Project_all_emtities and replace the following code-
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUXItLcovSAUy3IoS85JBjICi/KzU5BIFQwUFBaVYnWglI2Rl7qlFuYl5lUjqjGDqjPGrM4YqM0FW5piTo5CaV5JZkplajKTWBGamKRGKTaFqzYBizhmZeYkgPYnFJUhKzGDmmSOpwWGcuVJsLAA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [id = _t, Division = _t, #"Org." = _t, #"Project " = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"id", Int64.Type}, {"Division", type text}, {"Org.", type text}, {"Project ", type text}}),
#"Trimmed Text" = Table.TransformColumns(#"Changed Type",{{"Division", Text.Trim, type text}, {"Org.", Text.Trim, type text}}),
#"Filtered Rows" = Table.SelectRows(#"Trimmed Text", each ([#"Org."] = "All entities")),
#"Merged Queries" = Table.NestedJoin(#"Filtered Rows", {"Division", "Org."}, Organization, {"Division", "Custom"}, "Organization", JoinKind.LeftOuter),
#"Expanded Organization" = Table.ExpandTableColumn(#"Merged Queries", "Organization", {"Entity "}, {"Organization.Entity "}),
#"Reordered Columns" = Table.ReorderColumns(#"Expanded Organization",{"id", "Division", "Org.", "Organization.Entity ", "Project "}),
#"Renamed Columns" = Table.RenameColumns(#"Reordered Columns",{{"Organization.Entity ", "Entity"}})
in
#"Renamed Columns"
Step-4: Go to Advanced Editor for table Project_others and replace the following code-
let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUXItLcovSAUy3IoS85JBjICi/KzU5BIFQwUFBaVYnWglI2Rl7qlFuYl5lUjqjGDqjPGrM4YqM0FW5piTo5CaV5JZkplajKTWBGamKRGKTaFqzYBizhmZeYkgPYnFJUhKzGDmmSOpwWGcuVJsLAA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [id = _t, Division = _t, #"Org." = _t, #"Project " = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"id", Int64.Type}, {"Division", type text}, {"Org.", type text}, {"Project ", type text}}),
#"Trimmed Text" = Table.TransformColumns(#"Changed Type",{{"Division", Text.Trim, type text}, {"Org.", Text.Trim, type text}}),
#"Filtered Rows" = Table.SelectRows(#"Trimmed Text", each [#"Org."] <> "All entities"),
#"Merged Queries" = Table.NestedJoin(#"Filtered Rows", {"Division", "Org."}, Organization, {"Division", "Entity "}, "Organization", JoinKind.LeftOuter),
#"Expanded Organization" = Table.ExpandTableColumn(#"Merged Queries", "Organization", {"Entity "}, {"Organization.Entity "}),
#"Reordered Columns" = Table.ReorderColumns(#"Expanded Organization",{"id", "Division", "Org.", "Organization.Entity ", "Project "}),
#"Renamed Columns" = Table.RenameColumns(#"Reordered Columns",{{"Organization.Entity ", "Entity"}})
in
#"Renamed Columns"
Step-5: Create a new table By appending Project_all_emtities and Project_others with this below code-
let
Source = Table.Combine({Project_all_emtities, Project_others}),
#"Sorted Rows" = Table.Sort(Source,{{"Division", Order.Descending}, {"Project ", Order.Ascending}})
in
#"Sorted Rows"
Here is the final output-

Answer 2

I have changed the name of the columns, for the table Project. The function CROSSJOIN doesn't like to have duplicate columns names. Organization table is the same.
PROJECT TABLE
+------------+------------------+--------------+-----------+
| id_project | Division_project | Org. | Project |
+------------+------------------+--------------+-----------+
| 1 | Europe | France | Project 1 |
+------------+------------------+--------------+-----------+
| 2 | Europe | Germany | Project 2 |
+------------+------------------+--------------+-----------+
| 3 | Europe | Germany | Project 3 |
+------------+------------------+--------------+-----------+
| 4 | Europe | All entities | Project 4 |
+------------+------------------+--------------+-----------+
| 5 | Europe | All entities | Project 5 |
+------------+------------------+--------------+-----------+
| 6 | China | East | Project 6 |
+------------+------------------+--------------+-----------+
| 7 | China | All entities | Project 7 |
+------------+------------------+--------------+-----------+
The code uses two variables _directrel and _allentity, their outputs correspond to tables. Each one of those uses a CROSSJOIN surrounded by a FILTER statement. The first uses a direct relationship between divisions and entities. The second variable only uses divisions as a link and filtering the Project only for All Entities.
After the RETURN statement, there is a UNION statement, that makes a union between the tables, surrounded by SELECTCOLUMNS to control the order of the columns.
Merge =
VAR _directrel =
GROUPBY (
FILTER (
CROSSJOIN ( Organization, Project ),
[Division] = [Division_project]
&& [Entity] = [Org.]
),
[id_project],
[Division],
[Org.],
[Entity],
[Project]
)
VAR _allentity =
GROUPBY (
FILTER (
CROSSJOIN ( Organization, FILTER ( Project, [Org.] = "All Entities" ) ),
[Division] = [Division_project]
),
[id_project],
[Division],
[Org.],
[Entity],
[Project]
)
RETURN
SELECTCOLUMNS (
UNION ( _allentity, _directrel ),
"id", [id_project],
"Division", [Division],
"Org.", [Org.],
"Entity", [Entity],
"Project", [Project]
)
The output: