Open ehsteve opened 1 year ago
This sounds like a very elegant solution. For expanding fields, can be fieldname(expand)
or fieldname(*)
.
I plan on adding a capability to have an array shape be defined by a preceding field in the packet. this allows multiple variable length fields in the same packet. I was told a few instruments being developed at LASP do this with their packets. in this spec, that could be written out as fieldname(otherfield)
. if we do that maybe we should change fieldname(expand)
and array_shape=“expand”
to just fieldname(*)
and array_shape=“*”
so no one thinks there is a field called expand.
We should really put together a page on the site which describes this csv format, too
Leaving this open after #49 finished to implement fieldname(*)
I would suggest the following convention
So
uint(shape)
. This matches the programmatic way to define it and does not add too much complexity to the csv file.