Forum Discussion

Mike_Y_123's avatar
Mike_Y_123
Copper Contributor
Oct 17, 2023

Row limit for Excel power query with setting "connection only"

Hi Excel Forum,

I know the limit of rows in Excel is 1048576. Therefore a Power query, that loads to a worksheet, needs to be under this limit.

 

Is it possible to exceed the row limit of 1048576 in a power query that is a "Connection only" (Query 1)

 

Then have a new query (Query 2) that has it's source set to (Query 1) but has filters to limit the number of rows

 

Only Query 2 would be loaded to an Excel worksheet

  • johnsonmathias1's avatar
    johnsonmathias1
    Copper Contributor

    Mike_Y_123 

    I am pretty new to Excel Power Query, but i frequently transform data in multiples of 5 to 8 million rows,

    Last week i was attempting to create a New Merged Query

    A 9 million row query with a 5 million row query, in hindsight this was a poor decision, but it worked after waiting for 60+mins

     

    in my little experience Power Query does not seem to have a upper limit, but strangely its default table output can only be to a Excel sheet, thus limiting it to 1 Mil as a table, But its "To a Connection" does not have a upper limit

    Also there is a add in called Dax Studio, it lets you overcome the export to table limitation by saving your Query table output to a CSV file,

     

     

    • SergeiBaklan's avatar
      SergeiBaklan
      MVP

      johnsonmathias1 

      If you have millions of rows most probably you query SQL database. Or maybe OData endpoint. If something like this pay attention to query folding, that could increase the performance dramatically.

       

      If to load query into the grid it's limiting b grid size, i.e. bit more than 1M of rows. If load to data model on practice it's limited only by your PC memory.

       

      In general if you do nothing within Excel perhaps it's easier to use Power BI Desktop.

  • Mike_Y_123 

    On practice it's limited only by computer resources. Performance could be quite low with such size. Think about query folding if it is applicable in your case.

Resources