Microsoft Fabric Community Conference 2025, March 31 - April 2, Las Vegas, Nevada. Use code MSCUST for a $150 discount.
Register nowThe Power BI DataViz World Championships are on! With four chances to enter, you could win a spot in the LIVE Grand Finale in Las Vegas. Show off your skills.
Hi Team,
I have a table which shows representative, Region, Date, Items and No. of Units. I just want to look at the previous date units to for the date.
Say Example, For 2018-01-06 -->East-->Jones-->Pencil-->Previous Unit should be the unit is previous date for the same combination.
Regards,
Pradeep
Solved! Go to Solution.
If you want to do this as a calculated column you could use something like the following
Prev Units = var _Region = Table1[Region] var _Representative = Table1[Representative] var _item = Table1[Items] var _orderDate = Table1[OrderDate] var _prevDate = MAXX(FILTER(Table1, Table1[OrderDate] < _orderDate && Table1[Items] = _item && Table1[Region] = _Region && Table1[Representative] = _Representative), Table1[OrderDate]) var _result = LOOKUPVALUE(Table1[Units], [Region] , _Region, [Representative], _Representative, [Items], _item, [OrderDate] ,_prevDate) return _result
If you want to do this as a calculated column you could use something like the following
Prev Units = var _Region = Table1[Region] var _Representative = Table1[Representative] var _item = Table1[Items] var _orderDate = Table1[OrderDate] var _prevDate = MAXX(FILTER(Table1, Table1[OrderDate] < _orderDate && Table1[Items] = _item && Table1[Region] = _Region && Table1[Representative] = _Representative), Table1[OrderDate]) var _result = LOOKUPVALUE(Table1[Units], [Region] , _Region, [Representative], _Representative, [Items], _item, [OrderDate] ,_prevDate) return _result
@d_gosbell Thanks for this solution. I understand that this gets the 1-period lag record. Is it possible to modify the formula to get 2-period lag, 1-period lead, 2-period lead values etc.?
Going back or forward an abitrary number of periods would not be efficient using the lookup approach. Possibly something like the following by calculating a rank of the dates would work better for that.
Thank you d_gosbell. Fantastic solution that you have provided.
User | Count |
---|---|
129 | |
73 | |
70 | |
58 | |
53 |
User | Count |
---|---|
190 | |
97 | |
67 | |
63 | |
54 |