Forum Discussion
Index column
AlexisOlson I need a row identifier cause I am spliting a factTable[Field] to a new table, with schema VARCHAR (e.g. a|b|c / e|f|g|h) using PATHITEM, which is why I need a row identifier between factTbl and derivedFactTbl to create relationship (1:M).
Greg_Deckler has a great solution which works very well on a narrow table (n X 1) but not on a wide table (n X m). Nevertheless, thank you.
I managed to create a surrogateID by concatenating all the columns particpating in SUMMARIZECOLUMNS and then I created derivedTable with that SID.
What problem do you need an index column to solve? It doesn't seem like an index for a table this size is feasible but maybe there's another way to solve whatever issue you want an index for.
Greg_Deckler's approach is a neat use of CONCATENATEX and PATHITEM but won't scale. In Chris Webb's article, What Is The Maximum Length Of A Text Value In Power BI? , he mentions that
DAX functions such as ConcatenateX() put a limit on the length of the text values that they can work with at around 2.1 million characters
citing jeffrey_wang.
AlexisOlson I need a row identifier cause I am spliting a factTable[Field] to a new table, with schema VARCHAR (e.g. a|b|c / e|f|g|h) using PATHITEM, which is why I need a row identifier between factTbl and derivedFactTbl to create relationship (1:M).
Greg_Deckler has a great solution which works very well on a narrow table (n X 1) but not on a wide table (n X m). Nevertheless, thank you.
I managed to create a surrogateID by concatenating all the columns particpating in SUMMARIZECOLUMNS and then I created derivedTable with that SID.