Performance Analyser shows slow DAX query in New file but fast in Original File?

[THEG72]

Hi PBI Community,

I am trying to figure out why my DAX query is really slow when i have replicated my solution into a new file?

Fast Version

Here is the output from the performance analyser for the fast version

// DAX Query Speed 6154 FAST
DEFINE
  VAR __DS0FilterTable = 
    TREATAS({"2019"}, '+ C A L E N D A R'[Fiscal Year])

  VAR __DS0Core = 
    SUMMARIZECOLUMNS(
      'Income Statement Template'[Row Index],
      'Income Statement Template'[Income Statement Items],
      __DS0FilterTable,
      "Selected_Year_Actuals", 'Income Analysis'[Selected Year Actuals],
      "Selected_Year_Previous_Actuals", 'Income Analysis'[Selected Year Previous Actuals],
      "TY_vs_PY_Actuals", 'Income Analysis'[TY vs PY Actuals],
      "TY_vs_PY_Actuals__", 'Income Analysis'[TY vs PY Actuals %]
    )

  VAR __DS0PrimaryShowAll = 
    ADDMISSINGITEMS(
      'Income Statement Template'[Row Index],
      'Income Statement Template'[Income Statement Items],
      __DS0Core,
      'Income Statement Template'[Row Index],
      'Income Statement Template'[Income Statement Items],
      __DS0FilterTable
    )

EVALUATE
  GROUPBY(
    __DS0PrimaryShowAll,
    "MinTY_vs_PY_Actuals__", MINX(CURRENTGROUP(), [TY_vs_PY_Actuals__]),
    "MaxTY_vs_PY_Actuals__", MAXX(CURRENTGROUP(), [TY_vs_PY_Actuals__])
  )

EVALUATE
  TOPN(
    501,
    __DS0PrimaryShowAll,
    'Income Statement Template'[Row Index],
    1,
    'Income Statement Template'[Income Statement Items],
    1
  )

ORDER BY
  'Income Statement Template'[Row Index],
  'Income Statement Template'[Income Statement Items]

Here is the same Visual in a new file and the output of the performance is a lot slower

Slow DAX performance

Here is the output from the Performance Analyser

// DAX Query Speed is 24915 SLOW
DEFINE
  VAR __DS0FilterTable = 
    TREATAS({"2019"}, 'Calendar'[Fiscal Year])

  VAR __DS0Core = 
    SUMMARIZECOLUMNS(
      'Income Statement Template'[Row Index],
      'Income Statement Template'[Income Statement Items],
      __DS0FilterTable,
      "Selected_Year_Actuals", 'Performance Analysis'[Selected Year Actuals],
      "Selected_Year_Previous_Acutals", 'Performance Analysis'[Selected Year Previous Acutals],
      "TY_vs_PY_Actuals", 'Performance Analysis'[TY vs PY Actuals],
      "TY_V_PY_Actuals__", 'Performance Analysis'[TY V PY Actuals %]
    )

EVALUATE
  TOPN(
    501,
    ADDMISSINGITEMS(
      'Income Statement Template'[Row Index],
      'Income Statement Template'[Income Statement Items],
      __DS0Core,
      'Income Statement Template'[Row Index],
      'Income Statement Template'[Income Statement Items],
      __DS0FilterTable
    ),
    'Income Statement Template'[Row Index],
    1,
    'Income Statement Template'[Income Statement Items],
    1
  )

ORDER BY
  'Income Statement Template'[Row Index],
  'Income Statement Template'[Income Statement Items]

In the slow version, I can see that the Evaluate TOPN function seems to be running a lot earlier than the fast version.

Can anyone advise why this is the case and what I can do to improve the speed to how the original is running?

Its got me stumped?

[v-chuncz-msft]

@THEG72 

You may check if it is affected by the data model size.

[THEG72]

@v-chuncz-msft 

Its a smaller data model. Does the query information supplied give any indications...and the queries look different too when extracted from the performance analyser