calculate multiple filtered values from same column

[Anonymous]

I do not know if my previous post went through, it looks like it vanished, so here it is again.

I have an excerpt of a table:

week	product_id	stock	sales
1	b_40	17	325
2	b_40	19	300
1	f_3	5	532
2	f_3	23	425

I need to calculate the buy amount by the store. To calculate this i need to do Stock week X - stock week X-1, then add sales amount for week X.

as example: buys week 2 for product b_40:

19(stock week 2) - 17(stock week 1) = 2.

2+ 300(sales week 2) = 302.

so in week 2 the store bought 302 units of product b_40.

my problem is that I do not know how to calculate the 19-17 part. for every product and every week doing this calculation

I was thinking about making a new column with something like: NewColumn = [stock]-[stock-1] which obviously wouldn't work.

help would be most appreciated

[Jimmy801 · Accepted solution]

Hello @Anonymous 

in power query you can work with index-column and grouping. Here an example that shows you how to handle your puzzle

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUUqKNzEAUobmQMLYyFQpVidayQhJ3BIkbmAAFgepT4s3BpKmIGxsBFcNETUCESYgQ2IB", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [week = _t, product_id = _t, stock = _t, sales = _t]),
    #"Sorted Rows" = Table.Sort(Source,{{"product_id", Order.Ascending},{"week", Order.Ascending}}),
    #"Changed Type" = Table.TransformColumnTypes(#"Sorted Rows",{{"week", Int64.Type}, {"product_id", type text}, {"stock", Int64.Type}, {"sales", Int64.Type}}),
    #"Added Index" = Table.AddIndexColumn(#"Changed Type", "Index", 1, 1),
    #"Grouped Rows" = Table.Group
    (
        #"Added Index", 
        {"product_id"}, 
        {
            {
                "AllRows", 
                (tbl)=> Table.RemoveColumns(Table.AddColumn(tbl, "TotalBought", (add)=> let stockpreviousweek = try Table.SelectRows(tbl,  each [Index]= add[Index]-1)[stock]{0} otherwise 0 in add[stock]- stockpreviousweek + add[sales]), {"Index"})
            }
        }
    ),
    #"Expanded AllRows" = Table.ExpandTableColumn(#"Grouped Rows", "AllRows", {"week", "stock", "sales", "TotalBought"}, {"week", "stock", "sales", "TotalBought"})
in
    #"Expanded AllRows"

Copy paste this code to the advanced editor in a new blank query to see how the solution works.

If this post helps or solves your problem, please mark it as solution (to help other users find useful content and to acknowledge the work of users that helped you)
Kudoes are nice too

Have fun

Jimmy

[v-alq-msft · Accepted solution]

Hi, @Anonymous 

Based on your description, I created data to reproduce your scenario. The pbix file is attached in the end.

Table:

You may create a custom column with the following m codes.

let 
week = [Week],pid = [Product_id],
laststock = 
try
Table.SelectRows(#"Changed Type",each [Product_id]=pid and [Week]=week-1)[Stock]{0}
otherwise 
0
in 
[Sales]+[Stock]-laststock

Result:

Best Regards

Allan

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Anonymous]

hi, Thanks for the help.

This solution worked, albeit with a bit of naggling. powerBI complained about the "changed type". It was remedied when I manually did a step before this where i changed the type from whole numbers to whole numbers on one of the columns.

Also the solution took a really long time to figure out, on my pc it took from mondayafternoon to somewhere between 2-7AM. (the full tableis 3.5M rows long)

[Jimmy801 · Accepted solution]

Hello @Anonymous 

in power query you can work with index-column and grouping. Here an example that shows you how to handle your puzzle

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMlTSUUqKNzEAUobmQMLYyFQpVidayQhJ3BIkbmAAFgepT4s3BpKmIGxsBFcNETUCESYgQ2IB", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [week = _t, product_id = _t, stock = _t, sales = _t]),
    #"Sorted Rows" = Table.Sort(Source,{{"product_id", Order.Ascending},{"week", Order.Ascending}}),
    #"Changed Type" = Table.TransformColumnTypes(#"Sorted Rows",{{"week", Int64.Type}, {"product_id", type text}, {"stock", Int64.Type}, {"sales", Int64.Type}}),
    #"Added Index" = Table.AddIndexColumn(#"Changed Type", "Index", 1, 1),
    #"Grouped Rows" = Table.Group
    (
        #"Added Index", 
        {"product_id"}, 
        {
            {
                "AllRows", 
                (tbl)=> Table.RemoveColumns(Table.AddColumn(tbl, "TotalBought", (add)=> let stockpreviousweek = try Table.SelectRows(tbl,  each [Index]= add[Index]-1)[stock]{0} otherwise 0 in add[stock]- stockpreviousweek + add[sales]), {"Index"})
            }
        }
    ),
    #"Expanded AllRows" = Table.ExpandTableColumn(#"Grouped Rows", "AllRows", {"week", "stock", "sales", "TotalBought"}, {"week", "stock", "sales", "TotalBought"})
in
    #"Expanded AllRows"

Copy paste this code to the advanced editor in a new blank query to see how the solution works.

If this post helps or solves your problem, please mark it as solution (to help other users find useful content and to acknowledge the work of users that helped you)
Kudoes are nice too

Have fun

Jimmy