I’ve been working a lot with Microsoft Power BI the past month. I’m finding a lot of the same types of problems. Here, I’m going to describe how to rank values within a grouping.

### Problem Description

I have a CSV that looks like the following. I need to rank the VALUE column by product. I am assuming that the ID column provides something resembling an order that values are inserted into the database. If I don’t have this, I could easily use a timestamp column.

ID PRODUCT_KEY VALUE
1 B01JHMVG5O 160
2 B001GAOTSW 42
3 B001GAOTSW 150
4 B073H4VVC7 43
5 B001GAOTSW 77
6 B073H4VVC7 101
7 B073H4VVC7 155
8 B073H4VVC7 75
9 B01JHMVG5O 75
10 B01JHMVG5O 46
11 B001GAOTSW 106
12 B073H4VVC7 158
13 B073H4VVC7 117
14 B01JHMVG5O 161
15 B01JHMVG5O 139

The RANK column may be defined with the following DAX expression.

RANK = RANKX(FILTER(sales001,
sales001[PRODUCT_KEY] = EARLIER(sales001[PRODUCT_KEY])),
sales001[INVENTORY_VALUE], , ASC)


Now that we know the rank by product, I would like to know what the previous high value. I want to lookup a value, where the PRODUCT_KEY matches and the RANK is equal to RANK - 1.

PREV_HIGH_VALUE = LOOKUPVALUE(sales001[VALUE], sales001[PRODUCT_KEY], sales001[PRODUCT_KEY], sales001[RANK], sales001[RANK] - 1)


Finally, I want to know the growth from the previous high to the current value. This is a simple subtraction problem. We just need to make sure we check for blanks.

GROWTH_FROM_PREV_HIGH = sales001[VALUE] - IF(ISBLANK(sales001[PREV_HIGH_VALUE]), 0, sales001[PREV_HIGH_VALUE])


The final table should look like the following.

ID PRODUCT_KEY VALUE RANK PREV_HIGH_VALUE GROWTH_FROM_PREV_HIGH
1 B01JHMVG5O 160 4 139 21
2 B001GAOTSW 42 1   42
3 B001GAOTSW 150 4 106 44
4 B073H4VVC7 43 1   43
5 B001GAOTSW 77 2 42 35
6 B073H4VVC7 101 3 75 26
7 B073H4VVC7 155 5 117 38
8 B073H4VVC7 75 2 43 32
9 B01JHMVG5O 75 2 46 29
10 B01JHMVG5O 46 1   46
11 B001GAOTSW 106 3 77 29
12 B073H4VVC7 158 6 155 3
13 B073H4VVC7 117 4 101 16
14 B01JHMVG5O 161 5 160 1
15 B01JHMVG5O 139 3 75 64

That’s it! Have fun exploring your data!

### Notes

The sample CSV is available at the following location: sales001.csv.