Week over Week change for the same document

[Anonymous]

Hello guys!

I have a table called "OPEN POs" that looks like this:

WeekNum	Supplier	PO number	Payment term
33	A	158	30 days
33	C	105	60 days
33	Y	469	30 days
33	A	422	90 days
33	Y	789	75 days
33	U	156	120 days
33	Y	111	45 days
34	A	158	60 days
34	C	105	60 days
34	T	88	90 days
34	Z	156	120 days
34	U	777	120 days
34	B	879	45 days
34	Y	789	45 days
34	U	156	120 days

I'm trying to build a matrix like this:

Week 34	PO number	Supplier	From	To
	158	A	30 days	60 days
	789	Y	75 days	45 days

The idea is to show the "PO numbers" that had its "Payment term" changed from one week to another.

The variation always needs to be from the current week (that will be selected on the slicer) versus the previous week.

Can you guys please help me on this one?

@mahoneypat ideas? 😁 😬

Thank you for the support! 🙂

[amitchandak]

@Anonymous , refer my blog on Week over a week using week rank. But for that, you need a week and year. In case you only have a week. You can use week in place of week rank.

Power BI — WTD
https://medium.com/@amitchandak.1978/power-bi-wtd-questions-time-intelligence-4-5-98c30fab69d3
https://community.powerbi.com/t5/Community-Blog/Week-Is-Not-So-Weak-WTD-Last-WTD-and-This-Week-vs-Last-Week/ba-p/1051123

[mahoneypat]

Please try these two expressions for From and To.  They seem to get your desired result.

From = var prevweek = Max(Terms[WeekNum])-1

return CALCULATE(MIN(Terms[Payment term]), Terms[WeekNum] = prevweek)

To = var latestweek = Max(Terms[WeekNum])

return CALCULATE(MIN(Terms[Payment term]), Terms[WeekNum] = latestweek)

If this works for you, please mark it as the solution.  Kudos are appreciated too.  Please let me know if not.

Regards,

Pat

[lbendlin · Accepted solution]

What should happen in week number 1 ?

Anyway.  Add two calculated columns  (measures are not required here as there is no filter impact)

Last Week = 
var p='OPEN POs'[PO number]
var s='OPEN POs'[Supplier]
var w='OPEN POs'[WeekNum]
return calculate(max('OPEN POs'[Payment term]),all('OPEN POs'),'OPEN POs'[PO number]=p,'OPEN POs'[Supplier]=s,'OPEN POs'[WeekNum]=w-1)

Different = SWITCH(TRUE(),ISBLANK('OPEN POs'[Last Week]),"no",[Last Week]='OPEN POs'[Payment term],"no","yes")

Then filter your table to Different="yes"

[Anonymous]

Let me try this solution..

btw, week number one can be null.. no reference to bring any variance. 

[lbendlin]

that's not what I meant.  Is your report limited to a single year?  Or would you expect week 1 look back at week 52 of the prior year etc?

[Anonymous]

good catch.. at some point i'll have the same week for a subsequent year

[Greg_Deckler]

@Anonymous - You should be able to borrow the concept from MTBF. See my article on Mean Time Between Failure (MTBF) which uses EARLIER: http://community.powerbi.com/t5/Community-Blog/Mean-Time-Between-Failure-MTBF-and-Power-BI/ba-p/339586