Forum Discussion
Slicer to Exclude
Anonymous ,
Ok, I see.
In that case, I would create a [groupID] column by merging the [ID] and [DayID] fields in the unsummarised table ('table').
Then reference this table to create your summarised table ('summTable') by grouping on [ID], [DayID], [groupID], and SUM[Amount].
Send to model and relate table[groupId] to summTable[groupID]. Change relationship filter direction to BOTH.
Use table[Type] as your slicer field, summTable[Amount] (bins) as your X axis, and COUNTROWS(summTable) as your frequency measure on Y axis.
*NOTE* Ensure you understand the behaviour of this model setup. Using an unsummarised dimension to filter summarised data may give you unexpected results i.e. the single [Type]s selected in the slicer will output entire summarised groups, not just values specifically related to that [Type].
Pete
Hi Anonymous ,
The unsummarised table structure will still work fine for a histogram, just put only [ID]and [DayID] on the axis and use your measure for Values. Alternatively, you can right-click on fields in the Fields list and select 'New group' to bin your values however you like.
Perhaps I'm misunderstanding your use case?
Pete
Hi BA_Pete,
So there's 2 things I want to do.
1. I want to be able to filter with the summarised table with 'Type' using a slicer based of the original table, as showcased above.
2. Based on the sumarised table, I want to make a histogram with 'Amount' bins on the x-axis and frequncy on the y-axis ( which comes by placing the 'Amount' in the values section). I also want to be able to filter using 'Type' from the original table.
Anonymous ,
Ok, I see.
In that case, I would create a [groupID] column by merging the [ID] and [DayID] fields in the unsummarised table ('table').
Then reference this table to create your summarised table ('summTable') by grouping on [ID], [DayID], [groupID], and SUM[Amount].
Send to model and relate table[groupId] to summTable[groupID]. Change relationship filter direction to BOTH.
Use table[Type] as your slicer field, summTable[Amount] (bins) as your X axis, and COUNTROWS(summTable) as your frequency measure on Y axis.
*NOTE* Ensure you understand the behaviour of this model setup. Using an unsummarised dimension to filter summarised data may give you unexpected results i.e. the single [Type]s selected in the slicer will output entire summarised groups, not just values specifically related to that [Type].
Pete
Hi BA_Pete,
When you say merge ID and DayID is this right?
groupID= CONCATENATE('Table'[ID], 'Table'[DayID])And then form the summarised table usingsumTable= SUMMARIZE(table, table[ID],table[DayID],table[groupID],"Amount", sum(table[amount]))Is this right?Hi Anonymous ,
I would do both in Power Query myself:
// Call this idTable let Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSgWAjIOGoFKuDEDIxABJOcCFjkLABiiqQkBFIyBksZAQVMgYJucCFTGBmOaKoMkVoNIaqMkPYCBMyh6qKBQA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [ID = _t, DayID = _t, Amount = _t, Type = _t]), chgTypes = Table.TransformColumnTypes(Source,{{"ID", Int64.Type}, {"DayID", Int64.Type}, {"Amount", Int64.Type}, {"Type", type text}}), addGroupID = Table.AddColumn(chgTypes, "groupID", each Text.Combine({Text.From([ID], "en-GB"), Text.From([DayID], "en-GB")}, "-"), type text) in addGroupID // idTableSumm let Source = idTable, groupRows = Table.Group(Source, {"ID", "DayID", "groupID"}, {{"Amount", each List.Sum([Amount]), type nullable number}}) in groupRowsBut if you don't have access to PQ, then I would do the following DAX for the [groupID] field (just to make sure there's no bleed between the ID's):
groupID = [ID] & "-" & [DayID]Your SUMMARIZE DAX looks correct already.
Pete