I don't understand why a stock graph isn't categorizing by deposit
I have a Stock table called "S Stock" in which I have the column [tipodeposito] (which is the type of deposit) and [deposit] (which is the deposit per se). I created a Stock measure that only took into account some of the types of deposits and deposits. However, when I make a table in the report and categorize it by type of deposit, it shows the entire stock without differentiation. I have made this two ways (keep in mind [cantidad] in quantity): 1. Stock = CALCULATE(SUM('S Stock'[cantidad]), OR('S Stock'[tipodeposito] = "Consignado", OR( 'S Stock'[deposito] = "ALMACEN PYH", OR ('S Stock'[deposito] = "BODEGA PYH", 'S Stock'[deposito] = "Casa Big Sur")))) In this case, you should know all those examples of [deposito] are just a share of a [tipodeposito] that is called "Bodega" (which encompasses other deposits and doesn't intersect with "Consignado"). However, when I make the bar graph and try to differentiate the stock by [tipodeposito], this is how it appears: 2. In this case, I created two measures in order to better differentiate the stock I want to take into account from [tipodeposito] = "Consignado" and from [tipodeposito] = "Bodega". So: Stock Bodega = CALCULATE(SUM('S Stock'[cantidad]), OR('S Stock'[deposito] = "ALMACEN PYH", OR('S Stock'[deposito] = "BODEGA PYH", 'S Stock'[deposito] = "Casa Big Sur"))) Stock Consignado = CALCULATE(SUM('S Stock'[cantidad]), 'S Stock'[tipodeposito] IN { "Consignado" }) Later, I sum them up in the Stock measure: Stock = CALCULATE(Stock[Stock Bodega]) + CALCULATE(Stock[Stock Consignado]) And this is the graph I get (even weirder, as the "Bodega" has all the stock available and "Consignado" just a part of it, them being a simple column which categorizes all the stock in two categories): How can I have a graph in which the total stock is divided into these two categories, but in which "Bodega" only has those three deposits I mentiones ("ALMACEN PYH", "BODEGA PYH", "Casa Big Sur"), and without using filters in the table, just in the formula? Thanks!
Legends in a graph
Hi there. I am currently comparing time periods graphically. However, when I put the agent or self-management legend to divide the bar in two as in the first image, it does not allow me. Is there any way I can put these two features on the same bar and not split it apart (like in chart two)?Multiparent Graph, Recursiveness
Hi community, I am addressing the following issue, stemming from 2 observations concerning graph structures in DAX: The build in PATH function does not support several parents for one child. DAX is not recursive. This means that the implementation of graph exploration strategies starting with only e.g., an adjacency list is not possible in DAX My business problem is simple to understand: in a hierarchical company structure revenue of a child organization is recognized for several parents. A common application would be bonus agreements, where employees get bonus payments for revenues in child organizations. These bonus agreements cover then a completely internal structure which has nothing to do with the structure of the legal entities of the company, they rather are arbitrary. For these reasons I executed the entire graph search in Power Query and then handed over the result for visualization. The approach is outlined below. I have then several questions, where I would like your help: in the approach below, are there elements which can be replaced by features that are out of the box either in Power Query or DAX? If not, is there a pattern here, where necessarily recursive work is executed in Power Query and then handed over to the Vertipaq engine? Are there any performance concerns when doing so? Approach: We have as toy model a turnover table which specifies the turnover in the period and the child organizations. Organizations can have turnover or not (if they are pure umbrella organizations they do not). The real model has over 100 organizations and ideally, we explore the graph to get the dept of graph and do not put in this information beforehand. Therefore, the algorithm must be recursive. This is results in the following adjacency list: {{1,{3}},{2,{3}},{3,{4,5}},{4,{7}},{5,{}},{6,{4}},{7,{}}} Now some standard Depth First Search (DFS) is applied: DFS = (adjacencyList as list, visited as list, queue as list) => let result = if List.NonNullCount(queue) = 0 then visited else let v = Pop(queue){0}, restQueue = Pop(queue){1}, children = GetNeighbours(adjacencyList, v) in if List.Contains(visited, v) then @DFS(adjacencyList, visited, restQueue) else @DFS(adjacencyList, List.Combine ({ {v}, visited}), List.Combine({restQueue ,children})) in result All vertices are taken and the children are determined: GetDescendants = (adjacencyList as list, v as number) => let children = GetNeighbours(adjacencyList, v ), result = DFS(adjacencyList, {}, children) in result with GetNeighbours = (adjacencyList as list, v as number) => let selection = List.Select(adjacencyList, (x) => x{0} = v), result = if List.Count(selection) > 0 then selection{0}{1} else {} in result The output can be transformed into a table with the added feature that a vertex is its own child. let Vertices = Table.SelectColumns(Turnover, "Vertex"), Descendants = Table.AddColumn(Vertices, "Col1", each let v = [Vertex], children = GetNeighbours(TurnoverTree, [Vertex]), result = {v, GetDescendents(TurnoverTree, v)} in result), ls1 = Table.Column(Descendants, "Col1"), ls2 = List.Transform(ls1, (ls) => let v = ls{0}, n = List.Count(ls{1}), elementToAdd = if n > 0 then let generator = List.Repeat({v},n), zip = List.Zip({generator, ls{1}}) in zip else {} in elementToAdd), ls3 = List.Combine(ls2), ls4 = List.Transform(Table.Column(Vertices, "Vertex"), (x) => {x,x}), ls5 = List.Combine({ls3, ls4}), tab = #table( {"Vertex", "Child"}, ls5) in tab The rest is then straight forward. We can now assign now turnovers to the parents and have a completely basic unformatted visual like this (a nice visual is not the point of the question) Again, my key questions are: In the approach above, are there elements which can be replaced by features that are out of the box either in Power Query or DAX? If not, is there a pattern here, where necessarily recursive work is executed in Power Query and then handed over to the Vertipaq engine? Are there any performance concerns when doing so? Thank you for your help. Best regards Christian
Diversity report - Using Measure in Legend of graph
I am trying to build a Diversity report for my organization. A simple example of the data I work with: Candidate IDGenderOrg lvl1Org lvl 2Org lvl2ShortlistedInterviewedOffered 1 Male Sport Squash Ball Yes Yes Yes 2 Female Sport Squash Ball Yes Yes Yes 3 Unknown Sport Squash Ball Yes Yes Yes I added an excel file with the above and a power PI report here: https://wetransfer.com/downloads/e15fd9ac80e4da54360195bf5c5552c420221125151306/8f48a8bb5a0a74c98be0f1c5aa6788f120221125151323/c20328 I need to show the diversity data for each characteristic Broken down for each recruitment step (shortlisted, Interview, Offer). As we cannot use measures in the x Axis of a graph I started by Adding a new column (column1) to use as reference: with each recruitment step. Using Dax I then created the below measure: Measure test = VAR Short = CALCULATE( SUM( Table1[Candidate Sum] ), Table1[Shortlisted] ="Yes") VAR Inte = CALCULATE( SUM( Table1[Candidate Sum]), Table1[Interviewed] ="Yes") VAR Offe = CALCULATE( SUM (Table1[Candidate Sum]), Table1[Offered] = "Yes") RETURN SWITCH( SELECTEDVALUE( 'Recrtuiment steps'[Column1]), "Shortlisting", Short, "Interview", Inte , "Offer", Offe ) This returns the Diversity data perfectly https://ibb.co/mtdDFtn Now my issue is that I need to stop the graph from showing the legend data (Gender) when the value is less than 5. The problem is that the legend cannot accept a value. Adding a if less than 5 works in removing the data: less than 5 test = VAR Short = CALCULATE( SUM( Table1[Candidate Sum] ), Table1[Shortlisted] ="Yes") VAR Inte = CALCULATE( SUM( Table1[Candidate Sum]), Table1[Interviewed] ="Yes") VAR Offe = CALCULATE( SUM (Table1[Candidate Sum]), Table1[Offered] = "Yes") RETURN SWITCH( SELECTEDVALUE( 'Recrtuiment steps'[Column1]), "Shortlisting", IF(Short < 5 ,Blank() , Short) , "Interview", IF(Inte <5 , Blank() , Inte) , "Offer", IF( Offe < 5 , Blank() , Offe ) ) This DAX will remove the data completely if less than 5. Good start but now it looks like we don't have any small numbers instead of hiding them. https://ibb.co/SBYQQ1w What I need is for the legend to be amended to merge/group categories that are less than into a new "Hidden" value. So instead of (picturing the graph): Non Binary 3% - 3 Unknown 2% - 2 Male 50% - 50 Female 45% -45 shortlist I currently get: Male 50% - 50 Female 47% -45 shortlist I need: Hidden 5% Male 50% - 50 Female 45% -45 shortlist Is this possible, or am I not using the correct approach? My set of raw data is bigger but I only selected this to understand the logic. Thank you
