Currently, errors that happen when decoding a column are missing information about:
- What column that is
- What file we are reading
This context is available higher up in the call stack that the main error is generate. We can add additional context by adding wrapper errors.
This approach is described nicely in https://sabrinajewson.org/blog/errors.
Basically, want to do things like:
enum DecodeError {
NumberOutOfRange,
...
}
struct DecodeFieldError {
field_name: String,
field_id: i32,
offset: u64,
source: DecodeError,
}
fn decode(field: Field, data: Bytes) -> Result<_, DecodeFieldError> {
...
}
enum ReadFileErrorKind {
IO(ObjectStoreError),
Decode(DecodeFieldError),
}
struct ReadFileError {
path: Path,
kind: ReadFileErrorKind,
}
fn read_file(path: Path) -> Result<_, ReadFileError> {
let data = read_data(path)?;
for field in data.schema() {
decode(field, data).map_err(|err: DecodeFieldError| {
ReadFileError { path, kind: ReadFileErrorKind::Decode(err) }
});
}
}
And then the error stack would look like:
Error: failed to read file 'data.lance'
Caused by:
0: failed to read field 'age' (id: 10) for row offset 10
1: number out of range
Currently, errors that happen when decoding a column are missing information about:
This context is available higher up in the call stack that the main error is generate. We can add additional context by adding wrapper errors.
This approach is described nicely in https://sabrinajewson.org/blog/errors.
Basically, want to do things like:
And then the error stack would look like: