Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!
Hello,
I'm new to Power Query and DAX, trying to recreate some reports in it to improve efficiency in their generation. Having some trouble mimicking a few calculations performed in Excel otherwise.
One such calculation I'm struggling to recreate is an INDEX MATCH Array type value, which compares the TIMESTAMP of when agent's perform an action (working a ticket/ Closing a ticket) and the role they were in when that action took place (accounting for job role changes over the years).
I have Timestamp based data in various tables, showing when agents worked tickets, closed tickets, etc. In another separate table I have EmployeeHistory info, which lists out their roles, teams, start date, and end date information.
Below is an example of the dataset
Table: MASTER3_TIMETRACKING
MrID, mrTIMEDATE, mrTIMEUSER
1234, 05/05/2019, E0001
1244, 05/06/2019, E0023
Table: tblEmployeeHistory
EmployeeID, Role, Team, Start Date, End Date
E0001, Tier 1, Help Desk, 01/01/2017, 12/31/2018
E0001, Tier 2, Help Desk, 01/01/2019, 01/01/2099
E0023, Network Engineer, Networking, 01/01/2019, 01/01/2099
I'm trying to get a column added to the end of the time tracking table, for later pivoting, which will show the agent's role and team based on the timestamp of action.
In excel the array formula looks like this:
={IFERROR(INDEX(EmployeeHistory[Role],MATCH(1,(EmployeeHistory[EmployeeName]=[mrTIMEUSER)*([@Date]>=EmployeeHistory[Start Date])*([@Date]<=EmployeeHistory[End Date]),0)),"Invalid")}
I need to recreate this concept in DAX referencing the tables and columns themselves.
Is there a way to accomplish this?
Sincerely,
Kristopher
Solved! Go to Solution.
this would work in M
let Source = MASTER3_TIMETRACKING, #"Merged Queries" = Table.NestedJoin(Source, {"mrTIMEUSER"}, tblEmployeeHistory, {"EmployeeID"}, "Custom", JoinKind.LeftOuter), #"Expanded Changed Type" = Table.ExpandTableColumn(#"Merged Queries", "Custom", {"Role", "Start Date", "End Date"}, {"Role", "Start Date", "End Date"}), #"Filtered Rows" = Table.SelectRows(#"Expanded Changed Type", each [Start Date] <= [mrTIMEDATE] and [End Date] >= [mrTIMEDATE]), #"Removed Columns" = Table.RemoveColumns(#"Filtered Rows",{"Start Date", "End Date"}) in #"Removed Columns"
idea is following:
1) you merge the queries based on EmployeeID (only)
2) you expand the [Role], [Start Date] and [End Date] (at this point you will have more rows than in the beginning)
3) you filter the dates based on the [mrTIMEDATE] (you should have the same number of rows as in 1) now)
4) you remove unnecessary columns
try this code
Column = VAR __MrTIMEUSER = 'MASTER3_TIMETRACKING'[mrTIMEUSER] VAR __mrTIMEDATE = 'MASTER3_TIMETRACKING'[mrTIMEDATE] VAR __EmployeeHistory = FILTER ( 'tblEmployeeHistory', 'tblEmployeeHistory'[EmployeeID] = __MrTIMEUSER && 'tblEmployeeHistory'[Start Date] <= __mrTIMEDATE && 'tblEmployeeHistory'[End Date] >= __mrTIMEDATE ) RETURN CALCULATE ( FIRSTNONBLANK ( 'tblEmployeeHistory'[Role], TRUE () ), __EmployeeHistory )
Hey @Stachu ,
Apologies for my naivete and inexperience, with regard to Excel Power Query and DAX programming, but I have a quick question regarding the code you shared.
Any time I try and generate a custom column which includes a variable declaration, I get the following error...
Token Eof expected.
It isn't the first solution I've ran into this with, but I haven't been able to figure out how it should function. Any guidance in this regard would be most helpful!!
For reference:
VAR __MrTIMEUSER = 'MASTER3_TIMETRACKING'[mrTIMEUSER] VAR __mrTIMEDATE = 'MASTER3_TIMETRACKING'[mrTIMEDATE] VAR __EmployeeHistory = FILTER ( 'tblEmployeeHistory', 'tblEmployeeHistory'[EmployeeID] = __MrTIMEUSER && 'tblEmployeeHistory'[Start Date] <= __mrTIMEDATE && 'tblEmployeeHistory'[End Date] >= __mrTIMEDATE ) RETURN CALCULATE ( FIRSTNONBLANK ( 'tblEmployeeHistory'[Role], TRUE () ), __EmployeeHistory )Receive Eof Error
Sincerely,
Kristopher
I think there is a bit of confusion here
in Power Query you use M, DAX is used in the model itself (in Excel it was called Power Pivot to differentiate)
you asked for DAX solution in your post, the one I provided will work in the model
for M based solution it's a bit different approach, I can followup later this week
Thank you for the great information, @Stachu! Apologies for my inexperience and lack of knowledge. Yes, I was confusing the portion relating to PowerQuery with that of Power Pivot, and trying to write DAX in the custom column for M-Code.
I suppose that means though that the solution would work if I apply it to the data model in Power Pivot. I'll have to give that a shot later to find out.
Any additional info you would like to provide would be most appreciated, as I am still learning all about Power BI, and need to identify best practices to follow.
@Anonymous
If the problem is solved, can you mark the relevant post as a solution? That way other users can find answers more easily
this would work in M
let Source = MASTER3_TIMETRACKING, #"Merged Queries" = Table.NestedJoin(Source, {"mrTIMEUSER"}, tblEmployeeHistory, {"EmployeeID"}, "Custom", JoinKind.LeftOuter), #"Expanded Changed Type" = Table.ExpandTableColumn(#"Merged Queries", "Custom", {"Role", "Start Date", "End Date"}, {"Role", "Start Date", "End Date"}), #"Filtered Rows" = Table.SelectRows(#"Expanded Changed Type", each [Start Date] <= [mrTIMEDATE] and [End Date] >= [mrTIMEDATE]), #"Removed Columns" = Table.RemoveColumns(#"Filtered Rows",{"Start Date", "End Date"}) in #"Removed Columns"
idea is following:
1) you merge the queries based on EmployeeID (only)
2) you expand the [Role], [Start Date] and [End Date] (at this point you will have more rows than in the beginning)
3) you filter the dates based on the [mrTIMEDATE] (you should have the same number of rows as in 1) now)
4) you remove unnecessary columns