Data to be merged on Visual Level

srlabhe please change the name of table from ICPL3 to your table.

1. Model Assumption:

Date (Snapshot date used by slicer), ID, Name, idnamekey (your comparison key or concatenation)

You also have: 
Start Date slicer

End Date slicer

Both slicers filter the same Date column from your view or a Date dimension.

2. Capture Slicer values:

create these two measures:

Start Selected Date =
MIN ( 'Date'[Date] )

End Selected Date =
MAX ( 'Date'[Date] )

3. Base Lookup measures:

this verifies if a row exits

exists at start:

Exists at Start =
VAR _Start = [Start Selected Date]
RETURN
CALCULATE (
COUNTROWS ( ICPL3 ),
ICPL3[Date] = _Start
)

exists at end

Exists at End =
VAR _End = [End Selected Date]
RETURN
CALCULATE (
COUNTROWS ( ICPL3 ),
ICPL3[Date] = _End
)

4. Detect modification

Key At Start =
VAR _Start = [Start Selected Date]
RETURN
CALCULATE (
MAX ( ICPL3[idnamekey] ),
ICPL3[Date] = _Start
)

Key At Start =
VAR _Start = [Start Selected Date]
RETURN
CALCULATE (
MAX ( ICPL3[idnamekey] ),
ICPL3[Date] = _Start
)

5. Final measure - Required column "Change" 

Change Status :=
VAR _StartExists = [Exists at Start] > 0
VAR _EndExists = [Exists at End] > 0
VAR _StartKey = [Key At Start]
VAR _EndKey = [Key At End]
RETURN

SWITCH (
TRUE(),

-- Record added in End Snapshot
NOT _StartExists && _EndExists,
"New",

-- Record removed from End Snapshot
_StartExists && NOT _EndExists,
"Deleted",

-- Present in both but data changed
_StartExists && _EndExists && _StartKey <> _EndKey,
"Modify",

-- Present in both and unchanged
_StartExists && _EndExists,
"No Change",

BLANK()
)

6. Build the visual

Column for table:

ID, Name, idbamekey, Change Status

For Interactive & filterable Change column, do below steps:
1) Create disconnected date slicer tables

(If you already have date tables used elsewhere, create two tiny disconnected tables just for the slicers.)

Modeling → New table:

StartDate = CALENDAR( DATE(2015,1,1), DATE(2030,12,31) )
EndDate = CALENDAR( DATE(2015,1,1), DATE(2030,12,31) )

Or create them however you prefer.

2) Create simple ChangeType table (for slicer)

Modeling → New table:

ChangeType = DATATABLE("Change", STRING, { {"New"}, {"Deleted"}, {"No Change"}, {"Modify"} })

Put a slicer on ChangeType[Change].

3) Add the main measure: ChangeLabel

Create a measure in your model (adjust names to match your table/columns; I use YourView, SnapshotDate, ID, idnamekey — replace as needed):

ChangeLabel =
VAR vStart = SELECTEDVALUE( StartDate[Date] )
VAR vEnd = SELECTEDVALUE( EndDate[Date] )
VAR vID = SELECTEDVALUE( 'YourView'[ID] )

// Determine presence in Start/End snapshots
VAR inStart =
CALCULATE(
COUNTROWS( 'YourView' ),
ALL( 'YourView' ),
'YourView'[SnapshotDate] = vStart,
'YourView'[ID] = vID
) > 0

VAR inEnd =
CALCULATE(
COUNTROWS( 'YourView' ),
ALL( 'YourView' ),
'YourView'[SnapshotDate] = vEnd,
'YourView'[ID] = vID
) > 0

// Grab the idnamekey or comparable value to detect change
VAR startKey =
CALCULATE(
MAX( 'YourView'[idnamekey] ),
ALL( 'YourView' ),
'YourView'[SnapshotDate] = vStart,
'YourView'[ID] = vID
)

VAR endKey =
CALCULATE(
MAX( 'YourView'[idnamekey] ),
ALL( 'YourView' ),
'YourView'[SnapshotDate] = vEnd,
'YourView'[ID] = vID
)

RETURN
SWITCH(
TRUE(),
inEnd && NOT inStart, "New",
inStart && NOT inEnd, "Deleted",
inStart && inEnd && endKey = startKey, "No Change",
inStart && inEnd && endKey <> startKey, "Modify",
BLANK()
)

Notes:
>> ALL('YourView') removes the visual filter so the measure can explicitly target the SnapshotDate rows.
>> If idnamekey may be BLANK or duplicates exist, you can change the aggregation used (e.g., CONCATENATEX or other logic).
>> Replace 'YourView' and column names with your actual table & column names.

4) Add filter measure: IsInSelectedChange

Create this measure and use it as a visual-level filter (set to is 1) on your table visual.

IsInSelectedChange =
VAR selChange = SELECTEDVALUE( ChangeType[Change] ) // value selected in ChangeType slicer
VAR rowChange = [ChangeLabel] // the computed change label for the row
RETURN
IF(
ISBLANK(selChange) || ISBLANK(rowChange),
1, // if no change selected or row has no label, keep it visible
IF(rowChange = selChange, 1, 0)
)

Final Fix:

>> Put ID, Name, idnamekey (or whatever columns) into a table visual.
>> Add the ChangeLabel measure into the visual (so it shows as a column).
>> In the Visual → Filters area, add IsInSelectedChange and set it to is 1.
>> Now the ChangeType slicer behaves like a normal filter — selecting New or Modify will filter the table to those rows.
>> Show both Name_Start and Name_End if you want to display old vs new values — create them with CALCULATE(MAX(...), FILTER(...)) pulling from the Start and End snapshots like the logic used above.
>> If you want to show a friendly message if Start/End not selected, show a card with IF(OR(ISBLANK(vStart), ISBLANK(vEnd)), "Please pick Start and End", ...).
>> Make sure both StartDate/EndDate slicers are single-select (so the measures return a single date).

=================================================================
Did I answer your question? Mark my post as a solution! This will help others on the forum!

Appreciate your Kudos!!

Jaywant Thorat | MCT | Data Analytics Coach
Linkedin: https://www.linkedin.com/in/jaywantthorat/

Join #MissionPowerBIBharat = https://shorturl.at/5ViW9

#MissionPowerBIBharat
LIVE with Jaywant Thorat from 15 Dec 2025
8 Days | 8 Sessions | 1 hr daily | 100% Free

Hi,

Share the download link of the PBI file.

Hi srlabhe ,

Thanks for reaching out to the Microsoft fabric community forum.

The behavior you're seeing is expected with the current slicer selections and setup. When the Start Date is set to an earlier snapshot and the End Date to a later one where some names or IDs are missing, those records are correctly marked as Deleted. This shows the ChangeLabel logic is working as intended, since it checks for each ID's presence across the two selected snapshot dates rather than merging the data.

As long as each slicer is in List mode with single select, the SnapshotDate column is of Date type, and the table visual includes the ID field for row context, your results are accurate. The solution meets both the original requirement and the follow-up scenario, so no further changes are needed unless you want to add more filtering or formatting.

Please find the attached PBIX and Screenshort file for your reference.

Best Regards,
Tejaswi.
Community Support