Option skip unparse-able DBF fields

[irees]

hi @twpayne

I use your excellent libraries extensively in my work. Thank you so much for making my projects possible - I would not have been able to use Go otherwise.

I'd like to use go-shapefile to load various shapefiles in my system, and replace a shell script that uses ogr2ogr.

However, there's one significant behavior difference that is preventing me from doing this (without otherwise pre-processing the files). The default behavior of ogr2ogr is to ignore fields that fail to parse (setting the value to null), and continue processing additional records in the shapefile. Currently, go-shapefile stops processing and returns an error in this case.

To match the ogr2ogr behavior, I'd like to propose a new option in ReadDBFOptions called SkipBrokenFields, which when set simply ignores errors returned by ParseRecord.

I've included a test file (Natural Earth 10m Populated Places, which has ** as a value in an integer field and demonstrates the error) and tests that check both the expected error behavior and behavior when SkipBrokenFields is set, for both the normal reader and the Scanner reader.

I'm open to feedback and alternate approaches. Please let me know your thoughts.

[twpayne]

Thank you for both your kind words and this excellent PR full of tests :)

Yup, this PR makes total sense. I'll merge it now. Do you have more PRs in the pipeline? If not, I'll tag a new version of go-shapefile after merging so you have a clean version for your go.mod.

[irees]

@twpayne

Thank you for accepting my PR!

I'm incorporating go-shapefile into a couple different projects (some public repos, some private repos) and may have additional contributions when everything is in production, but no other requests at this time, so you are welcome to tag a new version.

[twpayne]

OK, I tagged v0.0.6 with your contribution. Thank you again!