Reading array of compound datasets

[ehayes2000]

I know this is similar to #217, but I still can't figure this thing out (I'm also a beginner). I have an h5 file with an attribute which holds an array of compound datasets:

   ATTRIBUTE "Coordinates" {
               DATATYPE  H5T_COMPOUND {
                  H5T_STD_I32LE "RowStart";
                  H5T_STD_I32LE "ColumnStart";
                  H5T_STD_I32LE "Width";
                  H5T_STD_I32LE "Height";
               }
               DATASPACE  SIMPLE { ( 1 ) / ( 1 ) }
               DATA {
               (0): {
                     88,
                     120,
                     796,
                     800
                  }
               }
            }

This is the h5dump of that file attribute

I've tried parsing this many different ways and seem to always get an array of large random numbers that change when I rerun the program indicating that I am getting some kind of address / indeterminate result.
I've tried both of these structs

#[derive(hdf5::H5Type, Debug, Clone, PartialEq)]
#[repr(C)]
pub struct Coordinates {
    a: i32,
    b: i32,
    c: i32,
    d: i32,
}
#[derive(hdf5::H5Type, Debug, Clone, PartialEq)]
#[repr(C)]
pub struct Coordinates {
    coords: [i32; 4]
}

I'm trying to parse my attribute like this

let values: Array1<Coordinates> = file.attr("path/to/Coordinates")?.read()?;

I'm stumped. Any help would be appreciated.

[mulimoen]

The names of the compound type needs to match what is given by the datatype. You can achieve this by either using the names directly:

#[derive(hdf5::H5Type, Debug, Clone, PartialEq)]
#[repr(C)]
pub struct Coordinates {
    RowStart: i32,
    ColumnStart: i32,
    Width: i32,
    Height: i32,
}

Or by using the rename functionality of the derive macro:

#[derive(hdf5::H5Type, Debug, Clone, PartialEq)]
#[repr(C)]
pub struct Coordinates {
    #[hdf5(rename = "RowStart")]
    a: i32,
    #[hdf5(rename = "ColumnStart")]
    b: i32,
    #[hdf5(rename = "Width")]
    c: i32,
    #[hdf5(rename = "Height")]
    d: i32,
}

[ehayes2000]

Wow... that was it. Thank you 😘