Updates to File Parsing #105
Conversation
srvasude
commented
Jul 25, 2024
srvasude
commented
- Push test files in to separate subdirectory (for each of the data types).
- Rename files to use _ naming everywhere.
- Update util_io to do CSV parsing for the observations. This allows for spaces in the values (which is what we see in the PCLean datasets + other real datasets).
- Improve error handling in file parsing.
cxx/util_io.cc
Outdated
| } | ||
| }, | ||
| schema.at(relname))); | ||
| } catch(const std::out_of_range& e) { |
There was a problem hiding this comment.
Can we check schema.contains(relname) instead of try/catch? Do you know if try/catch here will make it harder to run this inside Google eventually, given the prohibition on exceptions?
I've been thinking about whether it's worth it to implement more graceful error handling at this stage (maybe propagating error codes up to main) and using asserts only to check for invalid program state. Let's talk next week about what to do with that -- in the meantime, I'd kind of prefer to just stick with assert everywhere for consistency.
There was a problem hiding this comment.
Updated. Sticking with assert seems reasonable, since that is Google compatible. Do you mean using abseil Status and passing that around (to be Google compatible)?
There was a problem hiding this comment.
I was thinking abseil Status, but I haven't done much thinking/investigating other options (there's std::expected, but that's not until C++23)
