Days Out of Stock - DateDiff between rows with 0 values in middle

[Scott_HPD]

Hi All, 

I am trying to calculate how long we were out of stock of a product from an inventory snapshot table (actual data below)

In the screenshot above we sold our last unit of SKU1 on March 4th, we then received a new batch of product on May 5th.  This means we were out of stock for 62 days.  This can happen in multiple periods over the course of a single sku.   The above sample is already filtered down to a single sku from a more massive table of all inventory edits by date.  I can either calculate this in this table as an added column for each gap period or pull a total of time out of stock into a SKU reference table.  Any help on this would be much appreciated.

EDIT: Added Actual Table Data

Thank you

Scott

EntryDt	StockCode	OnHand	Days Out Of Stock
4/30/2019 0:00	SKU1	22
5/27/2019 0:00	SKU1	22
5/28/2019 0:00	SKU1	20
6/2/2019 0:00	SKU1	40
6/4/2019 0:00	SKU1	40
6/11/2019 0:00	SKU1	38
6/13/2019 0:00	SKU1	36
6/17/2019 0:00	SKU1	34
6/27/2019 0:00	SKU1	34
7/15/2019 0:00	SKU1	32
7/22/2019 0:00	SKU1	31
8/6/2019 0:00	SKU1	27
8/26/2019 0:00	SKU1	25
9/12/2019 0:00	SKU1	25
9/18/2019 0:00	SKU1	23
9/24/2019 0:00	SKU1	19
10/1/2019 0:00	SKU1	17
10/4/2019 0:00	SKU1	15
10/8/2019 0:00	SKU1	12
11/12/2019 0:00	SKU1	10
11/26/2019 0:00	SKU1	10
12/6/2019 0:00	SKU1	8
12/11/2019 0:00	SKU1	4
12/13/2019 0:00	SKU1	24
1/8/2020 0:00	SKU1	20
1/9/2020 0:00	SKU1	18
1/15/2020 0:00	SKU1	14
2/5/2020 0:00	SKU1	6
2/20/2020 0:00	SKU1	4
3/4/2020 0:00	SKU1	2
3/10/2020 0:00	SKU1	0
3/16/2020 0:00	SKU1	0
3/18/2020 0:00	SKU1	0
3/20/2020 0:00	SKU1	0
3/25/2020 0:00	SKU1	0
3/26/2020 0:00	SKU1	0
4/4/2020 0:00	SKU1	0
4/14/2020 0:00	SKU1	0
4/17/2020 0:00	SKU1	0
5/5/2020 0:00	SKU1	24	62
5/6/2020 0:00	SKU1	22
5/8/2020 0:00	SKU1	20
5/12/2020 0:00	SKU1	18
5/15/2020 0:00	SKU1	18
5/22/2020 0:00	SKU1	16
6/1/2020 0:00	SKU1	14
6/2/2020 0:00	SKU1	12
6/3/2020 0:00	SKU1	8
6/8/2020 0:00	SKU1	4
6/9/2020 0:00	SKU1	2
6/12/2020 0:00	SKU1	1
6/26/2020 0:00	SKU1	25
7/2/2020 0:00	SKU1	25
7/3/2020 0:00	SKU1	21
7/20/2020 0:00	SKU1	19
7/21/2020 0:00	SKU1	18
7/29/2020 0:00	SKU1	18
8/6/2020 0:00	SKU1	16
8/11/2020 0:00	SKU1	12
8/12/2020 0:00	SKU1	38
8/13/2020 0:00	SKU1	37
8/21/2020 0:00	SKU1	41
8/24/2020 0:00	SKU1	37
8/26/2020 0:00	SKU1	33
8/27/2020 0:00	SKU1	31
9/3/2020 0:00	SKU1	29
9/8/2020 0:00	SKU1	27
9/18/2020 0:00	SKU1	25
9/25/2020 0:00	SKU1	23
10/5/2020 0:00	SKU1	21
10/8/2020 0:00	SKU1	28
10/14/2020 0:00	SKU1	28
10/16/2020 0:00	SKU1	26
10/23/2020 0:00	SKU1	24
10/30/2020 0:00	SKU1	22
11/5/2020 0:00	SKU1	52

[Icey · Accepted solution]

Hi @Scott_HPD ,

Please check:

1. Add an index column.

2. Create calculated columns.

StartDate = 
VAR CurrentRow = [Index]
VAR PreviousRow =
    LOOKUPVALUE ( Scott[OnHand], Scott[Index], CurrentRow - 1 )
VAR NextRow =
    LOOKUPVALUE ( Scott[OnHand], Scott[Index], CurrentRow + 1 )
VAR FirstIndexofeachStock =
    CALCULATE ( MIN ( Scott[Index] ), ALLEXCEPT ( Scott, Scott[StockCode] ) )
VAR LastIndexofeachStock =
    CALCULATE ( MAX ( Scott[Index] ), ALLEXCEPT ( Scott, Scott[StockCode] ) )
VAR StartDate_ =
    IF (
        [Index] <> FirstIndexofeachStock
            && [Index] <> LastIndexofeachStock,
        IF ( [OnHand] <> 0 && PreviousRow <> 0 && NextRow = 0, [EntryDt] )
    )
RETURN
    StartDate_

Days Out Of Stock Column = 
VAR ModifiedStartDate =
    CALCULATE (
        MAX ( Scott[StartDate] ),
        FILTER (
            Scott,
            Scott[StockCode] = EARLIER ( Scott[StockCode] )
                && Scott[Index] <= EARLIER ( Scott[Index] )
        )
    )
VAR CurrentRow = [Index]
VAR PreviousRow =
    LOOKUPVALUE ( Scott[OnHand], Scott[Index], CurrentRow - 1 )
VAR NextRow =
    LOOKUPVALUE ( Scott[OnHand], Scott[Index], CurrentRow + 1 )
VAR FirstIndexofeachStock =
    CALCULATE ( MIN ( Scott[Index] ), ALLEXCEPT ( Scott, Scott[StockCode] ) )
VAR EndDate =
    IF (
        [Index] <> FirstIndexofeachStock,
        IF ( [OnHand] <> 0 && PreviousRow = 0 && NextRow <> 0, [EntryDt] )
    )
RETURN
    IF ( EndDate <> BLANK (), DATEDIFF ( ModifiedStartDate, EndDate, DAY ) )

Best regards

Icey

If this post helps, then consider Accepting it as the solution to help other members find it faster.