I'm trying to count and sum based on multiple columns. When I tried the "group by" function in transform data, it gives me a timeout error.
Below is an illustrative example of what I'm trying to do. In the real dataset, the number of columns is 30+ and the number of possible entries in each column is also large, resulting in many unique combinations.
I'm not sure if there are other functionalities in Power BI that can achieve this, please send help!
Have this:
Want this:
Im going to go ahead and guess that perhaps you want to sum all your columns without knowing how many of them you have, and this code can work on any number of columns, grouping the first two
let Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(Source, {"Column1", "Column2"}, "Attribute", "Value"),
#"Grouped Rows" = Table.Group(#"Unpivoted Other Columns", {"Column1", "Column2", "Attribute"}, {{"Count", each Table.RowCount(_), Int64.Type}, {"SUM", each List.Sum([Value]), type number}}),
#"Pivoted Column" = Table.Pivot(#"Grouped Rows", List.Distinct(#"Grouped Rows"[Attribute]), "Attribute", "SUM", List.Sum)
in #"Pivoted Column"
In Power Query, on the Tranform ribbon, click "Group By" and enter these settings.
Related
I am new to Power BI Desktop and Power BI Query. The left part is the data, and the right part is my desired output in Power BI. I tried to use the Matrix Report Visual to format my data like the table on the right, but it only allowed me to select the Product column in the row. I hope to get the desired output without creating separate tables for each quarter.
You could bring your data into powerquery, and use this code to transform the data. You'd have to modify it to show your actual table name (Table1) and the name of the columns as appropriate. It basically groups on index/product to find the most recent target sale, expands the data back, then pivots to get the sideways view
let Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Grouped Rows" = Table.Group(Source, {"index", "product"}, {
{"data", each _, type table },
{"most_recent_target_sale", each Table.Sort(_,{{"time", Order.Descending}}){0}[target_sale], type number}
}),
#"Expanded data" = Table.ExpandTableColumn(#"Grouped Rows", "data", {"time", "sale"}, {"time", "sale"}),
#"Changed Type" = Table.TransformColumnTypes(#"Expanded data",{{"index", type number}}),
#"Pivoted Column" = Table.Pivot(#"Changed Type", List.Distinct(#"Changed Type"[time]), "time", "sale", List.Sum)
in #"Pivoted Column"
Hello beautiful people,
Could someone please help me with my below request, (noting that I am working with Power Query Editor). So I need it to be done using creating conditional columns in power query maybe?, please help.
I need to group users in a table based on a category with showing their count in multiple different fields, As per the below example:
I need results to be:
Muuuuuuuuch Appreciated
In powerquery, try this
Click select period and name columns
Right click, unpivot other columns
Click select period, attribute and value columns
Right click, group ... new column name Count, operation Count rows
Click select attribute column. Transform ... pivot column... values column Count, Advanced options, do not aggregate
let Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(Source, {"Period", "Name"}, "Attribute", "Value"),
#"Grouped Rows" = Table.Group(#"Unpivoted Other Columns", {"Period", "Attribute", "Value"}, {{"Count", each Table.RowCount(_), type number}}),
#"Pivoted Column" = Table.Pivot(#"Grouped Rows", List.Distinct(#"Grouped Rows"[Attribute]), "Attribute", "Count", List.Sum)
in #"Pivoted Column"
Alternately,
Click select period and name columns
Right click, unpivot other columns
Right click name column and remove
Right click value column and duplicate
Click select attribute column .. Transform pivot column ... Values column:Values Copy
let Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(Source, {"Period", "Name"}, "Attribute", "Value"),
#"Removed Columns" = Table.RemoveColumns(#"Unpivoted Other Columns",{"Name"}),
#"Duplicated Column" = Table.DuplicateColumn(#"Removed Columns", "Value", "Value - Copy"),
#"Pivoted Column" = Table.Pivot(#"Duplicated Column", List.Distinct(#"Duplicated Column"[Attribute]), "Attribute", "Value - Copy", List.Count)
in #"Pivoted Column"
Neither one is going to put in a blank row for a combination that does not exist like Dec No. Thats more complicated if required
Hi need to add 1000+ calculated columns in power bi which provide the count per entry, for example.*means calculated columns
ID
RankCode
Count_RankCode*
RankAdvance
Count_RanAdvance*
1000
AAA
2
XYZ
2
1001
AAA
2
XYA
1
1002
AAB
1
XYZ
2
found the right way to count in power BI DAX:
COUNTROWS(FILTER('24Jun_1973',[rankCode]=earlier([rankCode])))
Requirement:
add 1000 columns that count rows in probably in one code using DAX
or create the 1000 count cloumn in power query M language (need it to be fast since raw date is 60gb).
As suggested by #smpa01, I was able to complete this task using the tabular editor. Just used the DAX script in tabular editor, put my all my measures in there since I was able to create all expressio in excels as it is just repeating then voila, 1000 measures added.
example:
Measure '24Jun_1973'[measure]=calculate(COUNTROWS(FILTER('28Jun_1973',[rankcode]='28Jun_1973'[rankcode])))
Measure '24Jun_1973'[measure2]=calculate(COUNTROWS(FILTER('28Jun_1973',[rankAdvance]='28Jun_1973'[rankAdvance])))
I have no idea why you would want another 1000 columns.
If you really want to though, in powerquery, you could unpivot, group and count, append the results to original data, then re-pivot. I don't know how fast it would be. I suspect not very.
let Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Changed Type" = Table.TransformColumnTypes(Source,{{"ID", Int64.Type}, {"RankCode", type text}, {"RankAdvance", type text}}),
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"ID"}, "Attribute", "Value"),
// group and count
#"Grouped Rows" = Table.Group(#"Unpivoted Other Columns", {"Attribute", "Value"}, {{"Count", each Table.RowCount(_), type number}}),
#"Duplicated Column" = Table.DuplicateColumn(#"Grouped Rows", "Attribute", "Attribute - Copy"),
#"Change column name" = Table.TransformColumns(#"Duplicated Column",{{"Attribute - Copy", each "count_" & _, type text}}),
// append back to original table, then repivot
#"Merged Queries" = Table.NestedJoin(#"Unpivoted Other Columns",{"Attribute", "Value"},#"Change column name",{"Attribute", "Value"},"Table2",JoinKind.LeftOuter),
#"Expanded Table1" = Table.ExpandTableColumn(#"Merged Queries", "Table2", {"Count", "Attribute - Copy"}, {"Count", "Attribute - Copy"}),
#"Removed Columns" = Table.RemoveColumns(#"Expanded Table1",{"Attribute", "Value"}),
#"Renamed Columns" = Table.RenameColumns(#"Removed Columns",{{"Count", "Value"}, {"Attribute - Copy", "Attribute"}}),
combined = #"Unpivoted Other Columns" & #"Renamed Columns",
#"Pivoted Column" = Table.Pivot(combined, List.Distinct(combined[Attribute]), "Attribute", "Value")
in #"Pivoted Column"
I need to sum all values in each row and display them in a calculated column. As I deal with lots of columns in lots of tables, adding something like
CalculatedColumn = 'public table_name'[column1] + 'public table_name'[column2] + ... + 'public table_name'[column528]
is really inefficient. Is there a shorter way of doing this?
Yes, there is. You should "Unpivot other columns" and then "Group By" using the Query Editor.
Suppose this dataset:
item;col1;col2;col3;col4;col5
apple;1;2;3;4;5
orange;1;2;3;5;8
banana;1;2;4;6;8
Load it up, and open the query editor.
Choose "Unpivot Other Columns":
You should now see this:
On the "Transform" tab in the ribbon, choose the leftmost "Group By" option. And fill out the dialog like so:
You should now have the wanted end result:
You could also skip the Group By step and let your visualization handle that.
PS. Should you need a few non-summed columns too I recommend either creating a duplicate dataset with the same source and either linking it to the original table with a relationship, or merging it so you get a final table with all wanted columns.
Footnote, this is the Power Query that is generated for you:
let
Source = Csv.Document(File.Contents("D:\Experiments\PowerBi\denormalized.csv"),[Delimiter=";", Columns=6, Encoding=1252, QuoteStyle=QuoteStyle.None]),
#"Promoted Headers" = Table.PromoteHeaders(Source, [PromoteAllScalars=true]),
#"Changed Type" = Table.TransformColumnTypes(#"Promoted Headers",{{"item", type text}, {"col1", Int64.Type}, {"col2", Int64.Type}, {"col3", Int64.Type}, {"col4", Int64.Type}, {"col5", Int64.Type}}),
#"Unpivoted Other Columns" = Table.UnpivotOtherColumns(#"Changed Type", {"item"}, "Attribute", "Value"),
#"Grouped Rows" = Table.Group(#"Unpivoted Other Columns", {"item"}, {{"SumCol", each List.Sum([Value]), type number}})
in
#"Grouped Rows"
I'm pretty new using queries and I still need to get on the topic properly.. :)
I'm editing a query in Power BI, and I created a new column that shouls show the division of each cell within a column by the total of another column, as %.
I wrote the following: = Table.AddColumn(#"Renamed Columns", "CONTACTED (CVR)", each [CONTACTED]/[LEADS]), but it divide each cell of the column Contacted by each cell of the column Leads.
Is there any way to do it?
Thanks in advance.
Cheers,
AS
In Power Query:
= Table.AddColumn(#"Renamed Columns", "CONTACTED (CVR)", each [CONTACTED]/List.Sum(#"Renamed Columns"[LEADS]))
This should do the trick, or get you going in the right direction.
You're looking for the Table.Group function aka Transform->Group By in the GUI. Detailed info here: https://msdn.microsoft.com/en-us/library/mt260774.aspx .
let
Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Country", type text}, {"Leads", Int64.Type}, {"Contacted", Int64.Type}, {"User", type text}}),
#"Grouped Rows" = Table.Group(#"Changed Type", {"Country"}, {{"Leads_Total", each List.Sum([Leads]), type number}, {"Contacted_Total", each List.Sum([Contacted]), type number}}),
#"Added Custom" = Table.AddColumn(#"Grouped Rows", "Percent", each [Contacted_Total]/[Leads_Total]),
#"Removed Columns" = Table.RemoveColumns(#"Added Custom",{"Leads_Total", "Contacted_Total"}),
#"Renamed Columns" = Table.RenameColumns(#"Removed Columns",{{"Country", "Country2"}}),
#"Result" = Table.Join( Source , "Country" , #"Renamed Columns" , "Country2" , JoinKind.LeftOuter ),
#"Removed Columns1" = Table.RemoveColumns(Result,{"Country2"})
in
#"Removed Columns1"
Hope it helps.
I think this is what you are asking; you need to make your question a MWE next time.
Drag the value (my value is called "Place") into the toolbar on the right hand side twice.
Right click the second value and select "Quick calc"
Make sure the "Summerize Value By" is Sum.
then select percent of grand total under "Show Value As".
that should do it.
Considering your condition per each country how many contacted out of leads,
that'll do:
let
Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
Types = Table.TransformColumnTypes(Source,{{"Country", type text}, {"Leads", Int64.Type}, {"Contacted", Int64.Type}, {"User", type text}}),
AddColumn = Table.AddColumn(Types, "Contacted (CVR)", (x) => x[Contacted]/List.Sum(Table.SelectRows(Types, (y)=> x[Country] = y[Country])[Leads]), Percentage.Type)
in
AddColumn