Return first non blank date per account per group

[ekaponkratova]

I have a dataset:

How can I get the first non-blank date per account?

[v-shex-msft · Accepted solution]

Hi @ekaponkratova ,

According to your snapshot, did you mean you want to get the first date for each continuous date range per account group?

If this is a case, does your table data group are stored with continuous date or random date collection?

If you mean continuous data, you can write calculate columns to use date as index to looping table records and mark tag. (is continuous)

isContinuous =
VAR previous =
    LOOKUPVALUE ( [Date], T3[id], [id], T3[Date], [Date] - 1 )
VAR next =
    LOOKUPVALUE ( [Date], T3[id], [id], T3[Date], [Date] + 1 )
RETURN
    IF (
        AND ( [Date] <> BLANK (), OR ( previous <> BLANK (), next <> BLANK () ) ),
        1,
        0
    )

Otherwise, you need to enter to query editor and add an index column for each account group.

isContinuous =
VAR previous =
    LOOKUPVALUE ( [Date], T3[id], [id], T3[Index], [Index] - 1 )
VAR next =
    LOOKUPVALUE ( [Date], T3[id], [id], T3[Index], [Index] + 1 )
RETURN
    IF (
        AND ( [Date] <> BLANK (), OR ( previous <> BLANK (), next <> BLANK () ) ),
        1,
        0
    )

After you finish creating IsContinuous column, you can simply write calculate column formula to use this field to lookup correspond records. (power bi data model does not contain column/row index to find out specific records)

First Continuous Date = 
IF (
    [isContinuous] = 1,
    VAR nearlest =
        CALCULATE (
            MAX ( T3[Index] ),
            FILTER (
                ALLSELECTED ( T3 ),
                [id] = AVERAGE ( T3[id] )
                    && [Index] < EARLIER ( T3[Index] )
                    && [isContinuous] = 0
            )
        )
    RETURN
        LOOKUPVALUE ( T3[Date], T3[id], [id], T3[Index], nearlest + 1 ),
    IF ( [isContinuous] = 0 && [Date] <> BLANK (), [Date] )
)

Regards,

Xiaoxin Sheng