Calculate last day with condition

[M4R10]

Dear all,

I have a data set where:
Items are audited, reporting 0 or 1 as a result.

But not all items are audited every day, the frequency is random. So some days, where the item was not audited, the result is reported as NA (not available)

See an example below for 3 items since March 7th until March 19th:

Audit date	07-03-2017	08-03-2017	09-03-2017	10-03-2017	11-03-2017	12-03-2017	13-03-2017	14-03-2017	15-03-2017	16-03-2017	17-03-2017	18-03-2017	19-03-2017
Item A	NA	NA	0	0	1	1	NA	1	0	NA	0	NA	NA
Item B	0	0	1	1	NA	1	NA	0	NA	0	0	0	0
Item C	1	1	1	0	1	1	1	NA	1	0	NA	0	0

What I'm trying to create are 5 DAX formulas:

DAX FORMULAS	Last audit date	Previous audit to last audit date	last audited "0" date	last audited "1" date	Consequtive days audited "0" since last audited "0" date
Item A	17-03-2017	15-03-2017	17-03-2017	14-03-2017	1
Item B	19-03-2017	18-03-2017	19-03-2017	12-03-2017	4
Item C	19-03-2017	18-03-2017	19-03-2017	15-03-2017	2

Where,
1. Last audit date = It is the last date different to "NA"

2. Previous audit to last audit date = It is the penultimate date different to "NA"

3. Last audited "0" date = It is the last date equal to "0"

4. Last audited "1" date = It is the last date equal to "1"

5. Consequtive days audited "0" since last audited "0" date = How many consequtive days were "0" since the last day equal to "0"

Anyone could help me?

thanks in advance!

regards

[Greg_Deckler]

First of all, I would unpivot your date columns to end up with something like:

Item,Audit date,Value

Item A,07-03-2017,NA

Item A,08-03-2017,NA

Item A,09-03-2017,0

...

At that point, most of your calculations should be relatively trivial along the lines of wrapping a MAX function for Audit date in a CALCULATE with an appropriate FILTER clause.

[M4R10]

thanks @Greg_Deckler

Actually, my tables are normalized, I have two tables:

- VISITS to register visit dates

- AUDIT to register the answers of each visit

According to your suggestion, I created 3 new measures:

LAST AUDIT DATE = CALCULATE(MAX('TBL_VISITS'[AUDIT DATES]);
FILTER('TBL_AUDIT';'TBL_AUDIT'[ANSWER]>=0))

LAST 0 AUDIT = CALCULATE(MAX('TBL_VISITS'[AUDIT DATES]);
FILTER('TBL_AUDIT';'TBL_AUDIT'[ANSWER]=0))

LAST 1 AUDIT = CALCULATE(MAX('TBL_VISITS'[AUDIT DATES]);
FILTER('TBL_AUDIT';'TBL_AUDIT'[ANSWER]=1))

However I still have an issue, the results of "audit 0" and "audit 1" are reported ONLY when it was 0 or 1 respectively:

In the following sample, the column "Last 0 Audit" should report 13-12-2016 for the 2nd and 3rd row (03-01-2017 and 16-02-2017):

Any idea?

thank you!!!

[v-sihou-msft · Accepted solution]

@M4R10

In this scenario, you apply the filter to limit the table context into "Answer = 0", the measure will not return result when the current context is "Answer =1". You may apply IF condition to combine the "last audit 0" and "last audit 1" into one measure.

To calculate the consecutive appearance, please refer to my expression in following thread:

http://community.powerbi.com/t5/Desktop/Count-last-actual-consecutive-sorted-rows/m-p/149683

Regards,