ADDCOLUMNS and actual values return (e.g. text string)

Greetings:

I have a question for the Community as follows:

Background:

I used an INTERSECT() to generate a single column table that is a return of distinct values of "Type A" that are common between Table 1 and Table 3.  Now I wish to generate an additional (extended) column using ADDCOLUMNS() to return all values of "Type B" from Table 2, which match the "Type A" values returned in the first column via INTERSECT() statement.  Note that Table 2 contains "Type A" and "Type B" values (so there is a granularity transition there.)

Since ADDCOLUMNS() is an iterator, I am able to wrap the ADDCOLUMNS() expression argument in CALCULATETABLE which in turn is wrapped by an Iterator function (e.g. COUNTROWS() to return e.g. an accurate count of "Type B" values matching each "Type A" value.  But, how might one go about returning the mathing "Type B" values (text) as opposed to -in this example- a count thereof?

Example Measure =

Part 1 - return single column table containing common "Type A" values btw Tables 1 and 3:

INTERSECT(DISTINCT(Table1[Type A Values]),DISTINCT(Table 3[Type A Values]))

Part 2 - return count of mathing "Type B" values (for each "Type A" value in the above resulting column):

ADDCOLUMNS(Part 1 expression),"Cnt of Matching Type B Values",COUNTROWS(CALCULATETABLE(DISTINCT(Table 2[Type B Values])

Entire expression:

ADDCOLUMNS(INTERSECT(DISTINCT(Table1[Type A Values]),DISTINCT(Table 3[Type B Values])),"Cnt of Matching Type B Values",COUNTROWS(CALCULATETABLE(DISTINCT(Table 2[Type B Values])

Again, the second column in the resulting virtual table ("Cnt of Mathing Type B Values") populates accurate counts of matching "Type B" values.  The question is how might one go about actually returning all of the mathing "Type B" values -for each "Type A" value, as opposed to merely their count?

Notes:

* I udnerstand same result for part 1 could be obtained via CALCULATETABLE(DISTINCT(Table 2[Type A Values]),RELATEDTABLE(Table 1),RELATEDTABLE(Table 2)) - let me know if you think this is a better solution and the reasons why.

* My problem in reality is a bit more complex in that it involves additional transitions in data granularity via additional lookup tables and fact tables but if I can get assistance with resolving the above example - that is how does one get ADDCOLUMNS() to return actual values which correspond to an adjacent column generated by it's <Table> argument- I should be able to apply the concept more broadly.

* I've tried IF(COUNTROWS(DISTINCT(Table 2[Type B Values]))>1,BLANK(),DISTINCT(Table 2[Type B Values])) but no cigar...

Let me know if you have any thoughts.  Thanks in advance!

Igor

Sample from actual data set (involves an additional set of values - say "Type C")

Needs ItemMathing EDLsMatching Rev.0 Dwgs