The .csv data exports are optimized for viewing in Excel, not optimized for programmatic scripting. It can be done, but it's does require nontrivial parsing.
Fields aren't guaranteed to have a particular absolute position, but they are at predictable relative locations. (For instance, if you have a large number of digital channels configured, that will increase the number of rows that the configuration section occupies, but the data section follows right after. The instruments are also in the same order left-to-right.)
If trying to parse out the MSO analog data programmatically, I'd go for a strategy like:
- Walk down rows until you find "Data" in the first column. This will skip over the Configuration sections (no matter how many rows they take up)
- Walk up one row, then walk right until you find the instrument you want. (Since you're interested in the "MIXED SIGNAL OSCILLOSCOPE", column zero is already the right place.)
- Walk down two rows, then right until you find "Analog"
- Walk down one row, then right one column. This is the "Timestamp" column header. Subsequent columns are for each channel, and the number will vary, as you have seen, based on number of enabled channels and sampling mode.
- You can then walk down processing each row.
The Read From Spreadsheet File LabVIEW VI might be useful here to read the whole .csv into a 2D array. I've not used Matlab, so I don't know how you'd accomplish that processing there.
