Handling arrays when request.security_lower_tf for 1 minute does not return values for each 1 min bar [pinescript v5]

I'm working with individual bars from lower time frames, e.g. working with 1 minute bars on a 5 minute chart. In principle request.security_lower_tf(syminfo.tickerid, '1',close) should return an array with 5 values for every bar on the 5 minute chart. However, 1 minute bars are often missing, even for highly liquid tickers. As an example, I've used the code below to plot the # number of 1 minute bars on a 5 minute chart for the ticker XLK. As you can see from the image, some 5 minute bars only have 3 or 4 one minute bars associated with them. The frequency of these events is even higher for medium and low volume tickers.

indicator("No of 1 min bars")

array_1min_close  = request.security_lower_tf(syminfo.tickerid, '1',close)   // gets the value of all 1 minute closes
num_1min_bars     = array.size(array_1min_close)                             // gets the size of the array, which is the total number of bars

plot(num_1min_bars,title="# of 1 min bars", color=color.yellow)              // plots the total number of 1 min bars

Here's the problem. Let's say the first of the expected 5 bars is missing. One would think that when a bar is missing that Pinescript would fill that position of the array with na, but instead it ignores that it is missing and fills it with the second 1 minute bar. This creates two coding issues:

1. not all arrays have uniform sizes, so errors often pop up when
   calling for index 3 or 4 from an expected 5 member array.
2. the placement of 1 minute bars is not uniform within the array, making it difficult to call the correct sub-bar e.g. in the case
   above bar #2 would be in array position 0 instead of the usual
   position 1

Is there a way to:

• produce these arrays in the expected manner, ie with all missing values filled with na in the proper position/order
• is it possible to achieve this in a manner that works for all time frames above 1 minute, e.g. the 5 minute chart always returning 5-membered arrays, a 30 min chart always returning 30-membered arrays, etc.

I have found a solution to issue 1 described above.

The solution is to add na values to the end of the array when request.security fails to do so, and to keep adding until the array has the expected number of elements. The code below compares the original code above (yellow plot) to the modified code where array.push() is used to add na values to the end of the array (fucschia):

indicator("No of 1 min bars")

// Original code where request.security enters elements without any modification
array_1min_close  = request.security_lower_tf(syminfo.tickerid, '1',close)   // gets the value of all 1 minute closes
num_1min_bars     = array.size(array_1min_close)                             // gets the size of the array, which is the total number of bars
plot(num_1min_bars,title="# of 1 min bars without push", color=color.yellow)  


// Modified code where na values are added to the end of the array to fill it to expected size (if request.security does not add the correct # of elements)
array_1min_close1  = request.security_lower_tf(syminfo.tickerid, '1',close)   
num_1min_bars1     = array.size(array_1min_close1) 
while barstate.isconfirmed and timeframe.isintraday and timeframe.multiplier > num_1min_bars1
    array.push(id=array_1min_close1, value=na)
    num_1min_bars1 := array.size(array_1min_close1)                            
plot(num_1min_bars1, title="# of 1 min bars w/solution 1", color=color.fuchsia)

The image below is taken from a 5 minute chart. request.security_lower_tf(syminfo.tickerid, '1',close) should theoretically have filled the array with 1 element per minute for a total of 5 elements in the array. The original code (yellow) shows this didn't happen in many instances, whereas there are always 5 array elements in the modified code (fuchsia).

Since this solution adds the missing array elements to the end of the array, it doesn't solve issue #2 above, where elements are entered into the array in the wrong position. There is no need to resolve this problem if your goal is to take the max or min of the array, to calculate an average, etc. But this remains an issue if you need to work with specific time-designated elements within the array, e.g. pick out the 4th 1 minute bar from the lower time frame. In principle, the solution to that problem would be to keep track of elapsed time and after every minute push na into the array if request.security_lower_tf did not push a value. However, I'm not sure if this can be accomplished because as far as I can tell, request.security_lower_tf only fills the array with lower time frame values AFTER the bar on the higher time frame is complete. If that is true, then it becomes impossible to retroactively insert na at time points where request.security_lower_tf did not.

Luckily this is the first coding language I've learned so I dont have much room to complain about how things SHOULD HAVE been formed a different way from the start. I just see the workarounds as being the normal way to punch through the walls to achieve the desired result. Sure, erasing the point in time where a candle would have formed but didnt bc there was no update to price/volume helps with reigning in the overuse of resources when possible (as well as other notable benefits to the overall product) but it also literally skews the data/calculations on scripts that use a Lookback(which is basically the entirety of the logic on the platform). But of course, theyll find some way to make it seem like any new issue that has popped up that requires a work-around was originally built with that intended purpose rather than it being an issue of foresight when creating the language...but nonetheless, we are here. Ill further add that I am very pleased with the product, even though it may be so only bc of my naivety to other coding languages in general. With almost every Scanner script I have built for the 1min TF, I have made use of this "desired effect" to enhance the end result (using it as a volatility filter to filter out alert triggers from coins that would otherwise take a lot more time to get the % gained (or lost) that I was looking for. Ok...enough jabbb.

One way you could address issue #2 in your inquiry would be to get a timestamp from all the printed bars and loop back through them to determine if the difference between a bars timestamp and the previous bars timestamp was the correct amount of time or if it showed that the time difference between the 2 timestamps is larger than the timespan of ONE of the bars. Hope this makes sense....and that it works but please do find me on TV (@chasinalts) and let me know if it worked or not (if you were going to give it a shit). If it doesnt work I have a few other ideas that I've implemented in some of my scanners that I know we could get working and get you on your way past this debacle. Peace brother....or sister.