How I can insert rows of zeros at beginning of a smaller matrix (X) until value of X(1,1) is less than or equal to first value encountered in column 1 of a longer matrix Y by n rows? n varies based data for Y. Resizing array X to be same size as Y.

Can you please help with how I can insert rows of zeros at the beginning of a smaller matrix (X) until the value of X(1,1) is less than or equal to the first value encountered in column 1 of a longer matrix Y by n rows? The row length n will always vary due to different data files imported for Y. The intent is to resize X to size of Y.

[KSSV]

confusing....can you give an example?

[Stephen23]

Thomas Gunther's "Answer" moved here:

I am working on an aerospace platform that has 3 devices recording telemetry data during a test flight. Two of the devices actually start recording before the third device is turned on to begin its recording and all data recording is stopped at the same time at end of the flight. The resultant telemetry data is extracted to create 3 separate and large CSV files. Column 1 of each file is the recorded time tag in seconds and the line count in the third CSV file is less than the other 2 CSV files so it's a smaller data file when it comes to Rows and Columns of data. The time between when the first 2 devices started their recording and the third device started recording could be seconds or several hundreds of seconds. When I analyze the results between instrument 1 and instrument 3, the data arrays won't be the same size and will be off by so many seconds (rows). My question is how can I compare the two arrays if they are different sizes and I want to start comparing when the two times are aligned. So I need to add rows of zeros to instrument 3 until the start time is aligned with the same point in time of instrument 1 and their data arrays are the same size. My ultimate goal is to write script that read in multiple csv files of any flight test telemetry data for data analysis, and time durations are not always the same.

[Stephen23]

@Thomas Gunther: is the same time period exactly the same? It would be highly unusual to have perfectly synchronized sample times, so even disregarding the different start time you will still have to consider the sampling times are all different. One easy solution is to interpolate the data to one common regular timebase, which would resolve both of these issues.

[Thomas Gunther]

Stephen, you are correct. The sampling times would likely be off by some small amount. I'm not sure if I follow if I interpolate to one common timebase (TB1), my issue would be resolved. The start time of INST-3 will still be off by TB1 + x seconds from INST-1 with its start time of TB1. Wouldn't I still have a smaller array for INST-3 than INST-1? Granted we would now know the exact time when INST-3 was turned on when the first data point taken occurs at the same point in time along TB1 when INST-1 is recording data. I still need to rectify the difference in array sizes for data analysis.

[Stephen23]

@Thomas Gunther: there are two separate topics to consider here:

1. finding a common epoch (i.e. synchronizing to the same time).
2. ensuring that the samples occur at the same time (e.g. interpolating, subsampling), if required.

Basically you have this:

TB1: S1   S2   S3   S4   S5    S6 ...
TB2:    S1     S2     S3     S5 ...
TB3:             S1 S2 S3 S4 S5 ...

Were they started by the same trigger, or do they use the same time signal? Do you have a common epoch for the sample times? Once you have that, then I see no reason why you could not interpolate, if required. If the samples are spread widely then interpolation may not be useful anyway. It depends on your data.