Search code examples
pythondatedaxadodbapi

DAX dates query in Python


I am running a DAX query within a Python script (using adodbapi) and would like to extract data from a table (called data) inbetween two dates (the field 'date_created'), however I am getting a syntax error.

tablestring = 'EVALUATE(CALCULATETABLE(Data,DATESBETWEEN(Data[date_created], 01/02/2021,04/02/2021)))'

df = pd.read_sql(tablestring, conn)

Solution

  • Here's a pattern that might work;

    • unpack the adoadapi results with a function to get it into a pandas frame
    • If you have some upstream process that handles this part using a conn object that read_sql can use, then you might only need to use the date() function on your DATESBETWEEN inputs; hard to see without the rest of your implementation code
    • sample code:

    `

        import adodbapi
        import pandas as pd
        import numpy as np
        
        workspace = 'your published workspace'
        report = 'report_name'
        
        sec = 'Integrated Security=SSPI'
        conn = adodbapi.connect(f"Provider=MSOLAP.8;Data Source='powerbi://api.powerbi.com/v1.0/myorg/{workspace}';Initial Catalog={report};{sec};")
        cur = conn.cursor()
         
        def get_df(data):
            ar = np.array(data.ado_results) # turn ado results into a numpy array
            df = pd.DataFrame(ar).transpose() # create a dataframe from the array
            df.columns = data.columnNames.keys() # set column names
            return df
        
        your_table_name = "Data"
        your_column_name = "date_created"
        
        DAX_Query = f'''
        DEFINE
          VAR FilteredTable = 
            FILTER(
               '{your_table_name}',
                AND(
                    '{your_table_name}'[{your_column_name}] >= DATE(2021, 1, 2),
                    '{your_table_name}'[{your_column_name}] < DATE(2022, 4, 2)
                  )
            )
        
        EVALUATE
          FilteredTable
        '''
         
        cur.execute(DAX_Query)
        data=cur.fetchall()
        demo_df = get_df(data)
    

    `