2026-07-22 10:24 UTC
@numeredevs@fosstodon.org @jwildeboer@social.wildeboer.net that depends on how the gap is represented. The original data was a CSV, where the state vector was a single field containing a JSON-style list.
I didn't actually know there were empty elements, so the following analysis is from memory how I did it.
In the video you can see that I first remove the open and close brackets, then split by comma and paste each element as a number. This results in a 6 element list of numbers.
If the field is empty, this would result in an empty list. When converting a list of lists into a 2-dimensional array, the number of columns in the result is tge same as the length of the longest list. Any elements that are shorter gets filled in with the fill element, which is 0 by default.
Thus, an empty element would become a row of 0's. If I had instead used the JSON parser to parse the list, an empty element would have resulted in an error.
Replies (1)
-
@numeredevs@fosstodon.org 2026-07-22 10:46
@loke@functional.cafe @jwildeboer@social.wildeboer.net Well, I might be wrong, but assuming you used https://cdn.gea.esac.esa.int/?prefix=Gaia/gfpr/Solar_system/sso_source/ as source, then there's at least one gap in file SsoSource_00.csv line 2614 (data row 2537), where no data is available for the state vector of the body. It's of course up to you and your design, how you want to handle such situations, but personally I'd consider to use either an empty field or NaN to signal the problem. Those "silent errors" can result in problems hidden for a long time.