Hi,
I have a flow which queries a SharePoint list and then create a csv out of it. List contains 100k items and GetItems action with filter condition would give me around 20k items. I need to process these 20k items inside an apply to each loop to convert date time formats . For this I am setting a variable to store the field value and then using a compose action to convert the date time format like below.

Before having this date conversion functionality inside apply to each loop , to speedup the process, I have set the Concurrency Control with Degree of Parallelism as 30 and my flow was running fast and could get the csv quickly. Here is the setting

But with the introduction of these additional actions (like set variable, compose etc.) inside apply to each I got warning message "Enable concurrency control for the apply to each loop and set its degree of parallelization to 1 when using a Set Variable action inside an apply to each loop."
With the concurrency ON and degree of parallelism set to 1, my flow is taking 3-4 hours to process 20k records . So question here is, with the concurrency setting on, should we not to use Set variable and compose actions.? Can some one clearly explain me what will happen if we set. And in case if we should not use the parallelism what are the best options to speed up this kind of process.?
-Purna