{"id":38859,"date":"2024-08-12T16:58:02","date_gmt":"2024-08-12T11:28:02","guid":{"rendered":"https:\/\/www.inogic.com\/blog\/?p=38859"},"modified":"2024-08-12T18:07:34","modified_gmt":"2024-08-12T12:37:34","slug":"use-parallel-branching-and-batching-to-accelerate-performance-in-power-automate-flow","status":"publish","type":"post","link":"https:\/\/www.inogic.com\/blog\/2024\/08\/use-parallel-branching-and-batching-to-accelerate-performance-in-power-automate-flow\/","title":{"rendered":"Use Parallel Branching and Batching to Accelerate Performance in Power Automate Flow"},"content":{"rendered":"

\"Use<\/p>\n

During a recent project, we faced a challenging requirement from a client: they needed to upload an Excel sheet containing multiple Opportunity product records and have these added to CRM using Power Automate Flow. Initially, the process of transferring a large volume of records from Excel to CRM was time-consuming and adversely affected system performance.<\/p>\n

To mitigate this issue, we implemented a solution to divide the Excel data into several batches and process them concurrently. This optimization significantly reduced processing time and improved overall system efficiency.<\/p>\n

The Excel file we are working with includes columns for Product Number, Product Name, and Quantity, outlined below:<\/p>\n

\"Use<\/p>\n

Step-by-Step Guide: Creating an Optimized Power Automate Flow in CRM<\/strong><\/h2>\n
    \n
  1. First, we created a File Type field on the Opportunity entity which takes the Excel file having Opportunity Product records. \"Use<\/a><\/li>\n
  2. We developed a Power Automate flow which triggers modification of the File type field. \"Use<\/a><\/li>\n
  3. In the next step, we added a “Download a file or an image” step and selected the File Type field, which provides the necessary file content. \"Use<\/a><\/li>\n
  4. Then we created a compose action to convert the file content into the string using the expressions string(body(‘Download_a_file_or_an_image’))<\/strong> \"Use<\/a><\/li>\n<\/ol>\n

    \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 This will give output as string of excel data. \"Use<\/a><\/p>\n

      \n
    1. Next, we created a compose Action with a line break to do this, simply press ENTER<\/strong>. \"Use<\/a><\/li>\n<\/ol>\n
        \n
      1. Next, we created a compose action to split each row of Excel by line break. For that, we used the expression split(outputs(‘Compose_CSV_content’),outputs(‘Compose_-_Line_Break’))<\/strong> \"Use<\/a><\/li>\n<\/ol>\n

        This will give us the array of rows. \"Use<\/a><\/p>\n

          \n
        1. Then we created compose Action in which we took out all the columns’ headings using the expression split(first(outputs(‘Compose_-_Split_Rows’)),’,’)<\/strong> \"Use<\/a><\/li>\n<\/ol>\n

          Below is the output from the above step: \"Use<\/a><\/p>\n

            \n
          1. Now we created a Select action to create an array of objects containing each record of Excel data.<\/li>\n<\/ol>\n

            skip(outputs(‘Compose_-_Split_Rows’),1):<\/strong> This expression skips the first row, which contains column headings.
            \noutputs(‘Compose_-_Column_Names’)?[0]:<\/strong> This retrieves the column heading at index 0; in our case, “Product Number<\/strong>” is at index 0.
            \noutputs(‘Compose_-_Column_Names’)?[1]:<\/strong> This retrieves the second column heading; in our case, “Product Name<\/strong>“.
            \ntrim(outputs(‘Compose_-_Column_Names’)?[2]:<\/strong> This trims any extra spaces from the last column heading, which is “Quantity” in our case.<\/p>\n

            \"Use<\/a>
            \nsplit(item(),’,’)?[0]:<\/strong> Retrieves the value of the first column.
            \nsplit(item(),’,’)?[1]:<\/strong> Retrieves the value of the second column.
            \nif(empty(split(item(),’,’)?[2]),null,trim(split(item(),’,’)?[2])):<\/strong> Checks if the value in the last column is blank; if so, it sets it to null; if not, it trims any extra spaces.<\/p>\n

            This step provides us with an array of objects, where each object represents a row value.<\/p>\n

            \"Use<\/a><\/p>\n

              \n
            1. Then we created a compose action to make batches of the array using the expression div(float(length(body(‘Select’))),5<\/span>)<\/strong><\/li>\n<\/ol>\n

              In this case, 5 batches are created, but you can adjust the number according to your needs. The expression calculates the batch size as a floating number based on the length of the array. \"Use<\/a><\/p>\n

                \n
              1. Then we created a compose action to make the batch size an integer and round it down using the expression int(sub(outputs(‘Batch_Size_in_float’), mod(outputs(‘Batch_Size_in_float’),1))) \"Use<\/a><\/strong><\/li>\n<\/ol>\n

                This expression converts the batch size from a floating-point number to an integer by rounding down. \"Use<\/a><\/p>\n

                  \n
                1. Then we added a parallel branch and added a Compose action in each branch with one batch. \"Use<\/a><\/li>\n<\/ol>\n

                  Every branch includes an equally divided array of objects using the expressions:<\/p>\n